conference_talk_with_qa AI powered creative strategy ·79 min ·Recorded Jun 2025

The #1 AI Creative Strategy That Helped This Agency Scale to 2k+ Ads Per Month

Mirella Crespi presents a framework for the "AI-augmented team of the future," arguing that Meta, TikTok, and YouTube algorithm changes now require concept-level creative diversity at a volume no human team can produce manually. She walks through her agency's process—concept development, AI creative sprints for rapid signal testing, and a multi-agent org chart spanning strategy, production, and post—before addressing AI authenticity concerns and compliance. The session closes with a Q&A and a Motion product demo by Evan Lee showing AI agents that generate creative insights, hook ideas, and UGC reworks from ad account data.

What's discussed, in order

00:00
Introduction: The AI-Augmented Team of the Future
00:19
About Ad Creative Academy & Creative Milkshake
01:18
The Next Era of Online Advertising
01:54
Platforms and Algorithms Are Changing Fast
02:24
AI-Powered Creative Strategy Framework
03:04
Why Creative Diversity Matters (Meta algorithm changes)
04:52
Concept Diversity + Structured Testing
05:38
Consumer Psychology Doesn't Change
06:01
Concept Development Framework
07:33
Creative Diversification Guide (Message, Formats, Execution, Placements, Best Practices)
08:06
Examples of Creative Diversity (Il Makiage, statics)
10:00
The Challenge and the AI-Powered Creative Sprint
12:43
Case Study: AI Creative Sprint
14:12
AI Agents vs. Agentic Teams
16:00
Multi-Agent AI Org Chart
19:22
How We Use AI in Creative Production
19:41
Blending AI + UGC
23:29
Augmenting Reality with AI
23:50
Elevating Hooks with AI
24:24
Creating B-Roll and Demos from Images
25:48
Augmenting Set Design
26:19
What AI Can't Do…Yet
27:15
Challenging "AI Can't Convey Human Emotion"
35:02
AI Has Forever Changed the Creative Process
36:19
The Audience Cares About the Story
38:10
Compliance Guidelines for AI Ads
39:57
Guidance for Team Leads and Team Members
42:43
Creative Milkshake and Ad Creative Academy CTA
43:42
Q&A Intro
46:30
Q&A: Lowest Viable Budget for AI Creative Sprints
49:07
Q&A: Tech Stack and AI Production Tools
51:37
Q&A: AI vs. Scripted UGC and Authenticity
56:45
Motion Product Demo: AI Agents & Tasks
1:04:07
Creative Strategy Flywheel
1:04:47
Briefing & Retrospective for Iteration

11 named frameworks

01 AI-Powered Creative Strategy
— Research-driven, strategic execution of high-quality creatives focused on concept and format variety. Visual: circular diagram with AI at the center surrounded by Creative Strategy, Production, Testing, Iteration, Development, Execution,…
02 Concept Diversity + Structured Testing
— ASC Campaign → multiple distinct Concepts → multiple Ads per concept → Iterations per ad. Introduced ~04:52.
03 Concept Development Framework
— Persona (who), Emotional Motivator (why), Benefits (what matters), Offer (what they buy) → Concepts. "Every ad creative is a hypothesis that needs to be tested." Introduced ~06:01.
04 Creative Diversification Guide
— 5-column matrix: Message, Formats, Execution, Placements, Best Practices. Introduced ~07:33.
05 Three Voices of Creative Diversity
— Customer Voice, Brand Voice, Expert Voice (applies to both video and static). Introduced ~08:14.
06 AI-Powered Creative Sprint
— Cyclical workflow: Concept Development → AI-Generated Image Variations → Human Instinct Gut Check → Test & Validate → Iterate on winners → Brief into full production. Introduced ~11:09.
07 AI Agent vs. Agentic Team
— Single-role AI agent vs. coordinated multi-agent team functioning as a creative pod. Introduced ~14:51.
08 Multi-Agent AI Org Chart
— AI Head of Strategy (Research, Ideation, Copywriting), AI Head of Production (Image, Video, Audio), AI Head of Post (Editing, QA), each with task lists and tool stacks. Introduced ~16:00.
09 Strategic Impact > Production Effort
— Shift creative team focus from output volume to strategic judgment. Introduced ~35:02.
10 Compliance Disclaimer Table
— Situation → Disclaimer text → FTC rationale (e.g., "Paid Actor. Not a Customer.", "AI-Generated Avatar.", "Simulated Imagery."). Introduced ~38:10.
11 Creative Strategy Flywheel (Motion)
— Research → Ideation → Briefing → Content Creation → Evaluation → Launch → Creative Analysis → Research. Introduced by Evan Lee ~1:04:07.

What's actually believed — in their own words

We are entering the next era of online advertising.

Crespi · 2025 · 01:18 (prediction) #

The volume of creative content needed to succeed has exploded beyond what any human team can manually produce.

Crespi · 2025 · 01:59 (observation) #

Meta now trains directly from every asset you upload.

Crespi · 2025 · 03:15 (technical) #

Small variations don't move the needle" under current algorithms.

Crespi · 2025 · 04:08 (observation) #

Platforms and algorithms change, but consumer psychology doesn't.

Crespi · 2025 · 05:51 (principle) #

Every ad creative is a hypothesis that needs to be tested.

Crespi · 2025 · 07:17 (principle) #

AI has forever changed the creative process.

Crespi · 2025 · ~35:02 (observation) #

The audience isn't focused on how content is created, they care about the stories we tell.

Crespi · 2025 · ~36:19 (opinion) #

AI creative sprints should be evaluated on signals rather than full spend thresholds (3–5x CPA, 2-week learning phase).

Crespi · 2025 · ~47:06 (tactical) #

The biggest differentiator between an average and a great AI creative is the editing…the post-production.

Crespi · 2025 · ~50:50 (opinion) #

AI video generation tools often cap at ~4 seconds per clip, requiring skilled editing to assemble.

Crespi · 2025 · ~51:20 (technical) #

Scripted UGC with paid actors is effectively a commercial; "authenticity" in paid social is largely a stylistic convention.

Crespi · 2025 · ~53:00 (opinion) #

The do's and don'ts pulled from the session

Do this
  • Crespi: Structure ASC campaigns with distinct concept folders, each containing diverse ad executions and iterations. 04:52 #
  • Crespi: Build concepts from Persona × Emotional Motivator × Benefits × Offer before briefing production. 06:01 #
  • Crespi: Diversify across Customer, Brand, and Expert voice — for both video and static. 08:14 #
  • Crespi: Run AI creative sprints with AI-generated static variations to validate hooks/headlines cheaply before committing to full video production. 47:06 #
  • Crespi: Map each team role's daily tasks and identify which are automatable by AI agents. 44:49 #
  • Crespi: Give team members dedicated paid AI learning time (Crespi gives her 30-person team a full paid AI day each month). 10:40 #
  • Crespi: Build personal AI agents to offload busy work and position yourself above them in the workflow. 42:30 #
  • Crespi: Blend AI generations (VO, Kling, ChatGPT images, Pika, Flux, Mirage, Arc Ads avatars) with UGC and rely on strong editing/post-production. 50:00 #
  • Crespi: Use compliance disclaimers: "Paid Actor. Not a Customer.", "Paid Actor Portrayal. Results may vary.", "AI-Generated Avatar.", "Simulated Imagery." 38:10 #
  • Lee: Use Motion AI agents to run account-wide tasks (social listening, persona ID, UGC sourcing) or per-creative tasks (new hook, new format, critique). 58:14 #
Don't do this
  • Crespi: Don't make small variations on a single creative expecting the algorithm to treat them as new tests. 04:08 #
  • Crespi: Don't use AI as a shortcut to "trick" the audience — viewers spot lazy AI. 36:41 #
  • Crespi: Don't wait for 3–5x CPA or full 2-week learning phase on AI sprint creatives; treat them as signal tests, not fully fleshed ads. 47:06 #
  • Crespi: Don't treat AI as the boss — keep human judgment as the steering wheel. 42:26 #
  • Crespi: Don't assume "UGC" means authentic — most paid-social UGC is scripted/paid and should be disclosed. 52:40 #

Numbers quoted in this talk

"The biggest advertisers…launch upwards of 50 to 100 creatives every week."
Crespi · 2025 · 02:18 #
AI video generation clip limits often ~4 seconds.
Crespi · 2025 · 51:25 #
Creative Milkshake team size: ~30 people.
Crespi · 2025 · 10:38 #
Motion offers a 14-day trial of its AI agents.
Lee · 2025 · 57:10 #

18 ads referenced

Show all 18 ads with extraction details
Ad #1 — Skincare UGC Example
unknown brand ·UGC, split-screen ·00:19
Duration shown in this video
2 seconds
Hook (first 3 sec)
A woman with a towel on her head looks at the camera, then a split screen shows her applying a product.
Product / pitch
A skincare or beauty product.
Key on-screen text
None used
Key spoken lines
None used
Visual style
UGC
CTA / offer (if shown)
None used
Narrative arc
None observable
Why shown in this video
To illustrate the types of creative production offered by the speaker's company, Creative Milkshake.
Speaker's take
The ad is part of a fast-moving collage showcasing different creative styles like strategy, production, testing, and iteration.
Ad #2 — Mobile Game Ad Example
unknown brand ·Mobile game ad ·00:26
Duration shown in this video
2 seconds
Hook (first 3 sec)
A 3D animated character runs through a colorful, obstacle-filled environment.
Product / pitch
A mobile game.
Key on-screen text
None used
Key spoken lines
None used
Visual style
Animated
CTA / offer (if shown)
None used
Narrative arc
None observable
Why shown in this video
To illustrate the types of creative production offered by the speaker's company, Creative Milkshake.
Speaker's take
The ad is part of a fast-moving collage showcasing different creative styles.
Ad #3 — IL MAKIAGE "Customer Voice" Ads
IL MAKIAGE ·UGC, Customer Voice ·08:06
Duration shown in this video
30 seconds
Hook (first 3 sec)
A collage of various women applying makeup, showing before-and-after results in a user-generated style.
Product / pitch
IL MAKIAGE makeup, likely foundation, for a flawless look.
Key on-screen text
"CUSTOMER'S VOICE"
Key spoken lines
None used
Visual style
UGC, mixed
CTA / offer (if shown)
None used
Narrative arc
A montage of user testimonials and application demos.
Why shown in this video
To demonstrate the "Customer Voice" pillar of creative diversity.
Speaker's take
"They have creatives that reach their consumers through the customer's voice. So the classic UGC videos, unboxings, demos, storytelling..."
Ad #4 — IL MAKIAGE "Brand Voice" Ads
IL MAKIAGE ·Polished, Brand Voice ·08:06
Duration shown in this video
30 seconds
Hook (first 3 sec)
A collage of high-end, polished visuals including close-ups of makeup products, swatches, and professional application shots.
Product / pitch
IL MAKIAGE makeup products.
Key on-screen text
"BRAND VOICE", "Free mystery gift"
Key spoken lines
None used
Visual style
High-fi, polished
CTA / offer (if shown)
"Free mystery gift"
Narrative arc
A montage of branded, high-quality product shots.
Why shown in this video
To demonstrate the "Brand Voice" pillar of creative diversity.
Speaker's take
"...creatives that come from the brand's voice..."
Ad #5 — IL MAKIAGE "Expert Voice" Ads
IL MAKIAGE ·Expert Voice, News/Report style ·08:06
Duration shown in this video
30 seconds
Hook (first 3 sec)
A collage of videos featuring experts, including a woman with red hair in a news anchor-style setup, discussing the product.
Product / pitch
IL MAKIAGE makeup products, endorsed by experts.
Key on-screen text
"EXPERT VOICE", "Collagen Loss in the Skin"
Key spoken lines
None used
Visual style
Mixed, professional
CTA / offer (if shown)
None used
Narrative arc
A montage of expert endorsements and scientific-style explanations.
Why shown in this video
To demonstrate the "Expert Voice" pillar of creative diversity.
Speaker's take
"...and creatives that come from this third or expert voice."
Ad #6 — Breakout Clips
Breakout Clips ·Video, AI Talking Head + UGC B-Roll ·20:57
Duration shown in this video
26 seconds
Hook (first 3 sec)
An AI-generated man speaks directly to the camera, saying, "Your boss wants a viral campaign in one hour? You need this tool."
Product / pitch
Breakout Clips is a tool with over 1,000 video templates to quickly create viral-worthy marketing campaigns without design skills.
Key on-screen text
"CREATE VIRAL-WORTHY CAMPAIGNS FAST", "1000+ VIDEO TEMPLATES", "MAKE ANY CONTENT POP", "IT LITERALLY TAKES 1 MINUTE", "NO DESIGN SKILLS NEEDED".
Key spoken lines
"Your boss wants a viral campaign in one hour? You need this tool. I used to waste hours jumping from different design apps and tools, but then I found Breakout Clips and it leveled up our campaigns... It literally takes one minute. Super easy, no design skills needed."
Visual style
Mixed (AI talking head, screen recordings, stock/UGC B-roll)
CTA / offer (if shown)
"Get started for FREE now on breakoutclips.com"
Narrative arc
Hook (problem) → Solution (product) → Features/Benefits → CTA.
Why shown in this video
As an example of blending AI-generated talking heads with user-generated B-roll footage.
Speaker's take
"So the first two ads that I'm going to play, these are AI talking heads, so AI avatars, blended with UGC B-roll."
Ad #7 — Perfect Tee
Unknown ·Video, AI Talking Head + UGC B-Roll ·21:25
Duration shown in this video
16 seconds
Hook (first 3 sec)
An AI-generated woman speaks to the camera, saying, "If a guy's wearing this on a date, he's coming home with me."
Product / pitch
A "perfect tee" for men, made of organic cotton and spandex, designed to enhance their physique.
Key on-screen text
"Date Night Game-Changer!", "They make his chest and arms look so good", "I'm not mad seeing him wear a different color every day".
Key spoken lines
"If a guy's wearing this on a date, he's coming home with me. My boyfriend owns seven of these perfect tees. They make his chest and arms look so good... They're made of 95% organic cotton and 5% spandex..."
Visual style
Mixed (AI talking head + UGC B-roll)
CTA / offer (if shown)
None used
Narrative arc
Hook (bold statement) → Personal story/social proof → Product features.
Why shown in this video
As a second example of blending AI-generated talking heads with UGC B-roll.
Speaker's take
"So classic AI talking heads with UGC B-roll, just super well edited."
Ad #8 — Wonderskin Lip Set
Wonderskin ·Video, UGC + AI B-Roll ·21:51
Duration shown in this video
39 seconds
Hook (first 3 sec)
An AI-generated voiceover says, "It looks like magic. It feels like nothing. It stays like it's part of you," over visuals of a lip product being applied and AI-generated B-roll.
Product / pitch
The Wonderskin Ultimate Lip Set, a long-lasting lip color that uses a scientific approach to infuse pigments into the lip layer.
Key on-screen text
"Welcome to the Lip Revolution", "Forget lipsticks", "This isn't makeup, it's science", "unbelievable results", "The hype is REAL".
Key spoken lines
"It looks like magic. It feels like nothing. It stays like it's part of you... This isn't makeup, it's science. Pigments infused directly into your top lip layer, sealed with our award-winning technology... We broke the internet and went viral for a reason."
Visual style
Mixed (UGC-style close-ups + AI-generated B-roll)
CTA / offer (if shown)
"Tap to transform your lips"
Narrative arc
Hook (intrigue) → Problem (old products) → Solution (new technology) → Benefits/Features → Social Proof → CTA.
Why shown in this video
To demonstrate a different blend of creative styles: using real UGC footage combined with AI-generated B-roll.
Speaker's take
"Now let's look at some different ways to think about and use AI... 'UGC' + AI B-Roll."
Ad #9 — HAVN Wavestopper Boxer Briefs
HAVN ·Video, AI + UGC B-Roll ·22:39
Duration shown in this video
48 seconds
Hook (first 3 sec)
An AI-generated image of a corn cob is shown with the text, "This is a shield to protect your [bleeped]."
Product / pitch
HAVN Wavestopper boxer briefs that use pure silver technology to block over 99.7% of wireless EMF radiation, protecting male reproductive health.
Key on-screen text
"This is a shield to protect your [bleeped]", "exposed to EMF radiation daily", "significantly reduce testosterone levels and sperm count", "Protection where it matters most".
Key spoken lines
"This is a shield to protect your [bleeped]... If you're watching this, it's because you're being exposed to EMF radiation daily without realizing the long-term consequences... HAVN's Wavestopper technology is different. Wavestopper boxer briefs block over 99.7% of wireless radiation with 360-degree coverage."
Visual style
Mixed (AI-generated visuals, stock footage, UGC-style clips)
CTA / offer (if shown)
"SHOP NOW"
Narrative arc
Hook (shock/intrigue) → Problem (EMF radiation) → Scientific proof → Solution (product) → Features/Benefits → CTA.
Why shown in this video
As another example of blending AI-generated visuals with UGC-style B-roll.
Speaker's take
"AI + 'UGC' B-Roll."
Ad #10 — AI-Augmented Reality Ads
Wonderskin, IL MAKIAGE ·Image (AI-generated) ·23:28
Duration shown in this video
21 seconds
Hook (first 3 sec)
N/A
Product / pitch
Makeup and beauty products from Wonderskin and IL MAKIAGE.
Key on-screen text
"WONDERSKIN", "IL MAKIAGE"
Key spoken lines
None used
Visual style
AI-generated, surreal
CTA / offer (if shown)
None used
Narrative arc
None observable
Why shown in this video
To illustrate how AI can be used to create surreal, scroll-stopping visuals by augmenting real-world cityscapes with giant product placements.
Speaker's take
"Another way we've been using AI is using it to augment reality... creating these really stunning visuals that are really scroll-stopping and help elevate creatives."
Ad #11 — AI-Elevated Hooks
Unknown ·Image (AI-generated) ·23:51
Duration shown in this video
32 seconds
Hook (first 3 sec)
N/A
Product / pitch
Various products related to bloating, baby clothes, and gaming.
Key on-screen text
"FEELING BLOATED?", "Still holding on to your baby's outgrown clothes?", "This is not fun after a long day of work"
Key spoken lines
None used
Visual style
AI-generated, surreal
CTA / offer (if shown)
None used
Narrative arc
None observable
Why shown in this video
To show how AI can create "out-of-the-ordinary" and visually engaging hooks for ads by representing problems in a surreal way.
Speaker's take
"And also elevating hooks. So AI is amazing to help with create out-of-the-ordinary things that really help us enhance the visuals of telling a story..."
Ad #12 — AI-Generated Hair Foil Demo
Unknown ·Video (AI-generated from still images) ·24:24
Duration shown in this video
19 seconds
Hook (first 3 sec)
A still image of a hairstylist applying foils transitions into an AI-generated video showing the process with different colored and patterned foils.
Product / pitch
Colorful, patterned hair foils for hairstylists.
Key on-screen text
"Original footage", "ChatGPT + Kling AI", "Uploaded References"
Key spoken lines
None used
Visual style
AI-generated video
CTA / offer (if shown)
None used
Narrative arc
Shows a before-and-after of how AI can turn static images into dynamic product demos.
Why shown in this video
To demonstrate how AI can create b-roll and product demos from existing images, improving production efficiency.
Speaker's take
"And creating b-roll and product demos from images... with AI, we're literally able to take images... and to create these really stunning, realistic demos just from some images."
Ad #13 — AI vs. Human Emotion Comparison
N/A ·Talking head, split-screen ·27:40
Duration shown in this video
58 seconds
Hook (first 3 sec)
A split screen shows two women, one real and one AI-generated, both expressing sadness and saying, "I will never forgive myself for this."
Product / pitch
N/A
Key on-screen text
None used
Key spoken lines
"I will never forgive myself for this." / "I don't understand why nobody told me about this before." / "You're really missing out if you haven't seen this."
Visual style
Lo-fi, selfie-style video
CTA / offer (if shown)
None used
Narrative arc
A series of side-by-side comparisons of real human actors and AI avatars expressing the same emotion and line.
Why shown in this video
To challenge the notion that AI cannot convey human emotion and to prompt a discussion about the future of AI in creative content.
Speaker's take
"But AI can't convey human emotion... I'll let you guess which one you think is AI and which one's human... It is crazy. It is disturbing... We are the ones that have to deal with this transition."
Ad #14 — Motion "Fire your ad agency"
Motion ·Video, Talking head ·59:01
Duration shown in this video
5 seconds
Hook (first 3 sec)
A woman looks directly at the camera. On-screen text appears: "YOU SHOULD FIRE YOUR AD AGENCY".
Product / pitch
Motion is a creative reporting and analytics platform for marketing teams.
Key on-screen text
"NEW TEST", "YOU SHOULD FIRE YOUR AD AGENCY"
Key spoken lines
"You should fire your ad agency."
Visual style
Polished, direct-to-camera, studio-like setting.
CTA / offer (if shown)
"Learn more"
Narrative arc
Hook is shown.
Why shown in this video
To demonstrate how Motion's AI Agents can be used to analyze and iterate on a specific ad.
Speaker's take
"So if I open up a creative, I'm actually able to see all the different tasks I can run... I would like for Mirella to re-imagine this ad in a different format."
Ad #15 — Motion "of inspiration"
Motion ·Video, UGC-style, Talking head ·1:00:01
Duration shown in this video
5 seconds
Hook (first 3 sec)
A man with a beard and glasses talks to the camera in a home/office setting.
Product / pitch
Motion is a tool that helps marketing managers make data-driven creative decisions and stop wasting ad budget.
Key on-screen text
"of inspiration"
Key spoken lines
"...of inspiration." (The full script is shown in the app's generated report: "My VP was on my back about ad ROI. As a marketing manager, I was burning through $100K monthly on ads that weren't converting.")
Visual style
UGC, lo-fi, direct-to-camera.
CTA / offer (if shown)
"Boost Your Paid Social Ads with Motion"
Narrative arc
Problem → Solution → CTA.
Why shown in this video
To demonstrate the output of an AI task that reworks an existing ad into a new, UGC-style format.
Speaker's take
"The outputs start to tell you like how do I rework this video into a UGC format with again different scripting options..."
Ad #16 — Simple Business USA
Simple Business USA ·Image ·1:04:16
Duration shown in this video
2 seconds
Hook (first 3 sec)
A man in a work setting looks at his phone with a concerned expression.
Product / pitch
Business insurance.
Key on-screen text
"Cut the cost of insuring your business."
Key spoken lines
None used.
Visual style
Polished, stock-style photography.
CTA / offer (if shown)
"Get quote"
Narrative arc
None observable.
Why shown in this video
To demonstrate the Motion Creative Research tool, which acts as a swipe file for ads.
Speaker's take
"What Creative Research is, is think about it as Pinterest for ads. So what you're doing is you're looking into the world at your competitors or brands you really admire..."
Ad #17 — AG1 by Athletic Greens (Ad Collection)
AG1 by Athletic Greens ·Image ·1:04:17
Duration shown in this video
10 seconds
Hook (first 3 sec)
A collection of clean, high-end product shots featuring the AG1 supplement.
Product / pitch
A daily nutritional supplement for energy, immune support, and overall health.
Key on-screen text
"All-day energy.", "Your Daily Powerhouse for Immune Support.", "For my Dad, the hero of my everyday life.", "What if your energy could last as long as your ambitions?"
Key spoken lines
None used.
Visual style
High-fi, polished product photography with consistent branding.
CTA / offer (if shown)
None used.
Narrative arc
None observable.
Why shown in this video
To demonstrate the Motion Creative Research and Brand Intel features for analyzing competitor ads.
Speaker's take
"So what you're doing is you're looking into the world at your competitors or brands you really admire to build out specific intel and boards."
Ad #18 — POP (Ad Collection)
POP ·Video and Image ·1:07:24
Duration shown in this video
10 seconds
Hook (first 3 sec)
A grid of colorful ads featuring skincare/beauty products, some with people using them, some as product shots.
Product / pitch
Skincare and beauty products.
Key on-screen text
"SELF-CARE SUNDAYS JUST GOT BETTER!", "50% OFF"
Key spoken lines
None used.
Visual style
A mix of UGC-style videos, polished product shots, and graphic templates.
CTA / offer (if shown)
"50% OFF"
Narrative arc
Varies by creative.
Why shown in this video
To demonstrate the "Creative Team" report in Motion, which uses color-coded scores (Hook, Watch, Click, Convert) to quickly assess creative performance.
Speaker's take
"The idea here is that they're built primarily for creative teams... we know the majority of people can speak in colors, especially the traffic light system... you're basically looking for your reds."

55 slides, in order

Show all 55 slides with full slide content
Slide #1 — The AI-augmented team of the future
title-only ·00:00 ·Play
Title / header text
The AI-augmented team of the future
Body content
• Mirella Crespi • Founder, Creative Milkshake • Director, Ad Creative Academy
Embedded data (charts/tables)
None used
Embedded examples
• Logo: Ad Creative Academy
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"Perfect. Let's do this. Hello, everyone. I am so excited to be back at Motion's Make Ads That Convert event... And today we're going to talk about the AI augmented team of the future."
Slide #2 — Ad Creative Academy & Creative Milkshake
2-column layout ·00:19 ·Play
Title / header text
None used
Body content
Left Column
• Logo: Ad Creative Academy • The World's First Academy for AI-Powered Creative Strategists • [Screenshot of a course page showing Seth Godin and Mirella Crespi]
Right Column
• Logo: creative milkshake • Circular diagram with a central clock icon, surrounded by concentric rings labeled: • CREATIVE STRATEGY • CREATIVE PRODUCTION • CREATIVE TESTING • CREATIVE ITERATION • CREATIVE DEVELOPMENT • CREATIVE EXECUTION • CREATIVE OPTIMIZATION • CREATIVE ANALYSIS • TikTok Creative Exchange Business Partner • Meta Business Partner • [Grid of 4 ad examples]
Embedded data (charts/tables)
None used
Embedded examples
• Ad Creative Academy course screenshots. • Creative Milkshake ad examples.
Annotations / visual emphasis
None used
Reveal state
The text in the circular diagram on the right animates in sequentially from 00:19 to 00:39.
Re-reference
None used
Speaker's framing
"So quickly about Ad Creative Academy, it's the world's first academy for AI powered creative strategists going live in July... And Creative Milkshake is a creative strategy and production studio that helps brands scale their paid social campaigns with direct response creatives."
Slide #3 — World's Biggest Advertisers
title-with-logo-grid ·00:49 ·Play
Title / header text
We have created and tested thousands of ads for the world's biggest advertisers
Body content
[A grid of 27 company logos]
Embedded data (charts/tables)
None used
Embedded examples
• Logos for: Lumen, Wix, Amazon, American Express, Wise, SpoiledChild, Unilever, ClearScore, Il Makiage, Elementor, Ubisoft, Johnson & Johnson, Loop, Deutsche Bank, Roman, Plaeo, SciPlay, Moon Active, Nectar, Playtika, Revolut, Vshred.
Annotations / visual emphasis
The phrase "biggest advertisers" is highlighted in yellow.
Reveal state
None used
Re-reference
None used
Speaker's framing
"At Creative Milkshake, over the past couple years, we have created and tested thousands of ads for the world's biggest advertisers."
Slide #4 — The Next Era of Online Advertising
title-only ·01:18 ·Play
Title / header text
We are entering the next era of online advertising
Body content
None used
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So, before we dive in, it's important to preface by saying that we are entering the next era of online advertising."
Slide #5 — Platforms and algorithms are changing fast
bullet list ·01:54 ·Play
Title / header text
Platforms and algorithms are changing fast
Body content
The volume of creative content needed to succeed has exploded beyond what any human team can manually produce.
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
The title "Platforms and algorithms are changing fast" is highlighted in a yellow box.
Reveal state
None used
Re-reference
None used
Speaker's framing
"These platforms and algorithms are changing very fast."
Slide #6 — AI-Powered Creative Strategy
image+text ·02:24 ·Play
Title / header text
AI-Powered Creative Strategy
Body content
Research-driven, strategic execution of high-quality creatives that focus on concept and formal variety.
Embedded data (charts/tables)
None used
Embedded examples
• A large, complex circular diagram on the right, showing multiple layers and segments related to creative strategy. The center shows a brain-like icon.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So this is what's going to happen now. It's an AI powered creative strategy era..."
Slide #7 — Why Creative Diversity Matters
bullet list ·03:04 ·Play
Title / header text
WHY CREATIVE DIVERSITY MATTERS
Body content
• Meta now trains directly from every asset you upload. • Meta's recent algorithm changes require a more structured creative development process. Small variations don't move the needle. Concepts, hooks, personas, formats, and execution styles must be different enough for the algorithm to identify them as different. • THE GOAL: CREATIVE VARIETY, VOLUME, AND VELOCITY
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
The line "Meta now trains directly from every asset you upload" is highlighted in a yellow box.
Reveal state
None used
Re-reference
None used
Speaker's framing
"So let's first talk about creative diversity."
Slide #8 — Concept Diversity + Structured Testing
hierarchy diagram ·04:52 ·Play
Title / header text
CONCEPT DIVERSITY + STRUCTURED TESTING
Body content
A flowchart showing: • Top box: ASC Campaign • Branches to three boxes: CONCEPT 1, CONCEPT 2, CONCEPT 3 (with text "Test Range on Diverse Concepts" on the arrows) • Each Concept box branches to three "AD" boxes (AD 1, AD 2, AD 3) • Each AD box has smaller boxes below it labeled "ITERATION 1", "ITERATION 2", "ITERATION 3".
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So moving forward, the best way to structure your creative strategy and think about how to develop your ads is imagining that they would be tested ideally, as recommended in an ASC campaign with different folders for each concept."
Slide #9 — Consumer Psychology
title-only ·05:38 ·Play
Title / header text
Platforms and algorithms change, but consumer psychology doesn't.
Body content
None used
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
The phrase "but consumer psychology doesn't" is highlighted in a yellow box.
Reveal state
None used
Re-reference
None used
Speaker's framing
"...to answer the question of what is a concept, we have to go back to the fundamentals of creative strategy and research and really getting inside the mind of your consumer. Because the truth is platforms and algorithms are going to change, but consumer psychology doesn't."
Slide #10 — Concept Development
diagram/template ·06:01 ·Play
Title / header text
CONCEPT DEVELOPMENT
Body content
A template with fields: • PERSONA -> Who they are • EMOTIONAL MOTIVATOR -> Why they buy • BENEFITS -> What matters to them • OFFER -> What they buy • Below these are three boxes: CONCEPT 1, CONCEPT 2, CONCEPT 3 • Bottom text: EVERY AD CREATIVE IS A HYPOTHESIS THAT NEEDS TO BE TESTED
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
The bottom line is highlighted in a yellow box.
Reveal state
None used
Re-reference
None used
Speaker's framing
"So when you are developing a concept, you want to identify the who, the why, and the what."
Slide #11 — Creative Diversification Guide
5-column table ·07:33 ·Play
Title / header text
CREATIVE STRATEGY + PRODUCTION
Body content
A table with 5 columns:
MESSAGE
Use different emotional motivators and barriers to reach new audiences. • Identify Human Desires: Motivation and barriers, Emotional Benefits, Structural Triggers • Product or Services: Functional Benefits, Product Usability, Demos • Promotions: Offers, Events
FORMATS
Contains video, static, animated graphics, carousels, and collections. • Core Formats: Static Image, Graphic Animation, Short Video, Long Video, Carousel, Carousel Ads • Bonus Formats: Show Ads (US only), Ads with product Tags • Dimensions: 1:1, 4:5, 9:16
EXECUTION
Use different creative execution styles to maximize differentiation. • Lo-Fi: UGC, Native/Mock Language, Creator-led • Hi-Fi/Polished: Product, Lifestyle, Illustration/Animation • Bonus: Sound first
PLACEMENTS
To make the most out of each channel's ecosystem, consider all placements. • Meta: Feed, Stories & Reels, Search, In-Stream, Audience Network • TikTok: In-Feed, Top View, Brand Takeover, Spark, Pangle (Audience Network) • YouTube: Display, Overlay, Skippable and Non, Bumper, Shorts
BEST PRACTICES
Use this checklist when developing and QA your ads. • Concept: Highlight brand in top 5 seconds, Use only real and audio hook, Show how your product or service solves a problem, Diversify talent, Strong CTA • Design: Optimize for placements, Use both static and video, Design for sound OFF, Work with sound ON, Use text overlays
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So this is a creative diversification guide that we like to refer to."
Slide #12 — Examples of Creative Diversity (Il Makiage)
3-column grid ·08:06, revisited 08:21 ·Play
Title / header text
EXAMPLES OF CREATIVE DIVERSITY
Body content
Column 1: CUSTOMER'S VOICE
3x3 grid of video thumbnails showing women applying makeup.
Column 2: BRAND VOICE
3x3 grid of video thumbnails showing product shots and makeup application.
Column 3: EXPERT VOICE
3x3 grid of video thumbnails showing various people, including a man in a lab coat, discussing the product.
Embedded data (charts/tables)
None used
Embedded examples
27 video thumbnails for the brand Il Makiage, categorized by voice.
Annotations / visual emphasis
None used
Reveal state
The slide is shown, then a series of short video clips from the thumbnails play from 08:22 to 08:58.
Re-reference
None used
Speaker's framing
"Let's look at a quick example of a brand that exercises creative diversity and volume and velocity very well. If you're familiar with Il Makiage, you know that they test hundreds if not thousands of creatives..."
Slide #13 — Examples of Creative Diversity (Various Brands)
3-column grid ·09:00 ·Play
Title / header text
EXAMPLES OF CREATIVE DIVERSITY
Body content
Column 1: BRAND VOICE
3x3 grid of video thumbnails showing polished product and lifestyle shots.
Column 2: CUSTOMER'S VOICE
3x3 grid of video thumbnails showing user-generated style content.
Column 3: EXPERT VOICE
3x3 grid of video thumbnails showing professional-looking individuals, including one in a lab coat and another on a stage.
Embedded data (charts/tables)
None used
Embedded examples
27 video thumbnails for various brands, categorized by voice.
Annotations / visual emphasis
None used
Reveal state
The slide is shown, then a series of short video clips from the thumbnails play from 09:03 to 09:39.
Re-reference
None used
Speaker's framing
"Let's look at other examples."
Slide #14 — Examples of Creative Diversity (Static Ads)
3-column grid ·09:40 ·Play
Title / header text
EXAMPLES OF CREATIVE DIVERSITY
Body content
Column 1: BRAND
2x2 grid of static ads with polished, branded visuals.
Column 2: NATIVE/USER
2x2 grid of static ads with a user-generated, "ugly" aesthetic (e.g., handwritten signs, social media comments).
Column 3: EXPERT
2x2 grid of static ads featuring press mentions (e.g., "Mirror" newspaper) and professional-looking visuals.
Embedded data (charts/tables)
None used
Embedded examples
12 static ad examples categorized by style.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"And this also translates to statics. So these would be examples of what brand, native user, and expert point of view would look like."
Slide #15 — Need Help Achieving Creative Diversity?
2-column layout ·10:00 ·Play
Title / header text
NEED HELP ACHIEVING CREATIVE DIVERSITY?
Body content
Left Column: PERSONA IDENTIFICATION
Screenshot of a detailed persona document.
Right Column: CONCEPT IDEATION
Screenshot of a concept ideation document. • [Image of the speaker, Mirella, with a "Get started" button]
Embedded data (charts/tables)
None used
Embedded examples
Screenshots of persona and concept documents.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"This is why I developed my AI agent together with Motion to help you achieve creative diversity."
Slide #16 — The Challenge & A Current Solution
bullet list ·10:41 ·Play
Title / header text
None used
Body content
The Challenge
Executing multiple concepts in so many formats to unlock winners requires production budget and time.
A Current Solution
AI-Powered Creative Sprint • AI Assets • Human Instinct • Data Validation
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
"The Challenge" and "A Current Solution" are highlighted in yellow boxes.
Reveal state
None used
Re-reference
None used
Speaker's framing
"Now, here's the challenge. Executing so many different concepts in so many different formats to unlock winners requires significant production budget and time. So a current solution that my team has found... is what we call the AI powered creative sprint..."
Slide #17 — AI Creative Sprint Diagram
flowchart/diagram ·11:09 ·Play
Title / header text
AI CREATIVE SPRINT
Body content
A circular flowchart showing the process: 1. Concept Development (Define Hypotheses) 2. Test & Validate (Run tests) 3. Brief Production in Post-Production (Scale what resonates, ditch what doesn't. Brief into production what gains traction.) 4. Iterate on static ads that gain traction 5. Test & Validate (Run tests) 6. Human Instinct / Gut Check 7. AI Generated Image Variations • The cycle repeats.
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So this is what it looks like. We sit down in a brainstorm with our clients or just with our team and we start filling out that concept map that I showed you..."
Slide #18 — Creative Strategy + Production Example
image+video grid ·12:43 ·Play
Title / header text
CREATIVE STRATEGY + PRODUCTION
Body content
• Left side: A grid of 4 AI-generated static images showing messy clothes. One is starred as the winner. • Right side: A video showing a woman overwhelmed by clothes, which then transitions to her using an app to sell them. • Text below: Winner guides briefing for videos, more images, animated ads, carousel - it informs casting, visuals, etc.
Embedded data (charts/tables)
None used
Embedded examples
• 4 static AI-generated images. • 1 user-generated style video.
Annotations / visual emphasis
A yellow arrow points from the static images to the video.
Reveal state
The video on the right plays from 13:13 to 13:15.
Re-reference
None used
Speaker's framing
"Here's an example of what this would look like."
Slide #19 — AI Agents > Agentic Teams
title-only ·13:12, revisited 14:12 ·Play
Title / header text
AI Agents > Agentic Teams
Body content
None used
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"Okay, let's kind of switch tabs and talk about AI agents and agentic teams."
Slide #20 — AI Agent vs. Agentic Team Definitions
bullet list ·14:51 ·Play
Title / header text
None used
Body content
AI Agent
Like one team member — e.g., a strategist who takes a brief and writes a creative angle, or a video editor who turns a storyboard into a polished cut. It works well on its own but sticks to its lane.
Agentic Team
Like your entire creative pod — a strategist, copywriter, producer, editor, and project manager working together to deliver a campaign. Each agent plays a role, passes work back and forth, adjusts based on feedback, and drives toward the final creative deliverable with minimal human involvement.
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So first, let's talk about the difference between an AI agent and an agentic team..."
Slide #21 — Multi-Agent AI Org Chart
organizational chart/diagram ·16:00 ·Play
Title / header text
MULTI-AGENT AI ORG CHART
Body content
A flowchart showing three main sections: 1. **AI Head of Strategy**: "Defines audience, angles and writes scripts." • Branches to: Research, Ideation, Copywriting (Individual Agents) • Each agent has a list of tasks (e.g., "Competitor Research", "Ad Account Audit", "Generate Hooks"). • Bottom section "Agent Tools" lists logos like ChatGPT, Perplexity, GigaBrain, Reddit, Foreplay, Motion. 2. **AI Head of Production**: "Sources talent and assets." • Branches to: Image, Video, Audio (Individual Agents) • Each agent has a list of tasks (e.g., "Generate Product Photography", "AI Avatar Creation", "Voiceover"). • Bottom section "Agent Tools" lists logos like Midjourney, DALL-E, Leonardo, Runway, Pika, ElevenLabs, Descript. 3. **AI Head of Post**: "Compiles edits and checks for quality." • Branches to: Editing, QA (Individual Agents) • Each agent has a list of tasks (e.g., "Generate Captions", "Check for Brand Safety"). • Bottom section "Agent Tools" lists logos like Adobe Premiere, After Effects, CapCut, Final Cut Pro, Kdenlive.
Embedded data (charts/tables)
None used
Embedded examples
Logos of various AI and creative software tools.
Annotations / visual emphasis
The header "AI-AUGMENTED CREATIVE TEAM" is highlighted.
Reveal state
None used
Re-reference
None used
Speaker's framing
"So let me share with you a little sneak peek into what our multi-agent AI org chart looks like to inspire yours as well."
Slide #22 — How we are using AI in Creative Production
title-only ·19:22 ·Play
Title / header text
How we are using AI in Creative Production
Body content
None used
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
The phrase "Creative Production" is underlined in yellow.
Reveal state
None used
Re-reference
None used
Speaker's framing
"All right, let's switch tabs again and let's talk about how we are using AI in creative production."
Slide #23 — Blending AI + "UGC"
4-column grid of video examples ·19:41 ·Play
Title / header text
Blending AI + "UGC"
Body content
• Column 1: AI Talking Heads + "UGC" B-Roll • Column 2: AI Talking Heads + "UGC" B-Roll • Column 3: "UGC" + AI B-Roll • Column 4: AI + "UGC" B-Roll
Embedded data (charts/tables)
None used
Embedded examples
Four video examples are shown and played sequentially.
Annotations / visual emphasis
None used
Reveal state
The videos play one after another from 19:57 to 20:45.
Re-reference
None used
Speaker's framing
"So one of the main things we're doing is blending AI with UGC."
Slide #24 — Augmenting Reality
2x2 grid of images ·23:29 ·Play
Title / header text
Augmenting Reality
Body content
Four images showing surreal, AI-generated scenes in real-world locations like Times Square, featuring giant product placements and brand-themed objects.
Embedded data (charts/tables)
None used
Embedded examples
• Image 1: Giant lipsticks and mascaras as skyscrapers in a city. • Image 2: A billboard in Times Square replaced with a different ad. • Image 3: A large hot air balloon with the "IL MAKIAGE" logo floating down a European street, with smaller balloons parachuting down. • Image 4: A giant phone screen displaying an ad, held by a couple in a busy city square.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"Another way we've been using AI is using it to augment reality."
Slide #25 — Elevating Hooks
3-column grid of images ·23:50 ·Play
Title / header text
Elevating Hooks
Body content
Three sets of before-and-after style images showing how AI can visually enhance the opening hook of an ad.
Embedded data (charts/tables)
None used
Embedded examples
• Image 1: A woman feeling bloated, with AI visuals representing the feeling. • Image 2: A woman overwhelmed by clothes, with AI visuals exaggerating the scene. • Image 3: A product being cut by a knife, and a woman in a fantasy-style outfit with a sword.
Annotations / visual emphasis
None used
Reveal state
The images within the slide animate and change from 23:51 to 24:22.
Re-reference
None used
Speaker's framing
"And also elevating hooks. So AI is amazing to help us create out of the ordinary things..."
Slide #26 — Creating b-roll and product demos from images
image grid ·24:24 ·Play
Title / header text
Creating b-roll and product demos from images
Body content
• Left image: "Original footage" showing a hairstylist applying foils. • Middle image: "ChatGPT + Kling AI" showing an AI-generated video of a hairstylist applying different colored foils. • Right images: "Uploaded References" showing the different colored foil products.
Embedded data (charts/tables)
None used
Embedded examples
Original footage, AI-generated video, and reference product images.
Annotations / visual emphasis
None used
Reveal state
The middle video animates from 24:35 to 24:56.
Re-reference
None used
Speaker's framing
"And then creating B-roll and product demos from images."
Slide #27 — Augmenting set design
screenshot-with-annotations ·25:48 ·Play
Title / header text
Augmenting set design
Body content
A screenshot of a video editing software (CapCut). It shows a feature where the background of a video clip is being replaced with an AI-generated set design.
Embedded data (charts/tables)
None used
Embedded examples
Screenshot of CapCut interface.
Annotations / visual emphasis
None used
Reveal state
The AI generation process is shown animating within the screenshot from 25:55 to 26:03.
Re-reference
None used
Speaker's framing
"We also use AI a lot to augment set design."
Slide #28 — What AI can't do...yet.
4-column grid of video examples ·26:19 ·Play
Title / header text
What AI can't do...yet.
Body content
Four columns, each with a 3x3 grid of video thumbnails showing various human-centric ad concepts like comedy skits, couple interactions, and product demos.
Embedded data (charts/tables)
None used
Embedded examples
Multiple video thumbnails.
Annotations / visual emphasis
None used
Reveal state
A montage of the video clips plays from 26:20 to 27:13.
Re-reference
None used
Speaker's framing
"Now here's what AI cannot do yet."
Slide #29 — "But AI can't convey human emotion"
title-only ·27:15 ·Play
Title / header text
None used
Body content
A blue speech bubble with the text: "But AI can't convey human emotion"
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So, just to challenge the conversation and explore kind of the biggest objection that I see at the moment, which is the talk about like authenticity and human emotion, but AI cannot convey human emotion."
Slide #30 — Discussion Questions
bullet list ·28:46 ·Play
Title / header text
None used
Body content
• How do you feel watching these as a human? • What is the difference between using a paid actor to read a script and using an AI avatar? • What's the difference between a good AI ad and a bad AI ad? • Will content creators be replaced by AI? • Will all content we consume be AI?
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"After watching this, how do you feel watching these as a human?"
Slide #31 — AI has forever changed the creative process.
bullet list with diagram ·35:02 ·Play
Title / header text
AI has forever changed the creative process.
Body content
• [Diagram: Strategic Impact > Production Effort] • AI can scale output, remix references, and generate endless variations. But it cannot tell us what will resonate. It cannot understand the nuance of what moves an audience. • Judgement. Taste. Risk. Humor. Weirdness. Cultural Sensitivity. • These are the things only you as a creative human can bring to the work. • The tools may change, but the essence of what we do, telling stories that connect, that make people feel, remains unchanged.
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...because AI has forever changed the creative process."
Slide #32 — The Audience Isn't Focused on How Content is Created
bullet list ·36:19 ·Play
Title / header text
None used
Body content
• The audience isn't focused on how the content is created, they care about the stories we tell, the problems your product or service solves, and how it makes them feel. • When AI is poorly used or when it's used with the intent to "trick" someone, no one will respond well to that. If we use AI as a shortcut, the audience will spot it. • If we use it with care, to augment ideas and elevate the stories we tell, they feel it too. When AI is used well, in bold, creative, funny, or beautiful ways, it's well received. • Enforcement is patchy, the trust is on us as creative teams to use AI ethically.
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"And the truth is the audience isn't really focused on how the content is created."
Slide #33 — Compliance Guidelines
table ·38:10 ·Play
Title / header text
Compliance Guidelines
Body content
A table with three columns:
SITUATION IN THE AD
• Actor plays a "customer" and offers an opinion or result • Actor dramatizes a scenario (no testimonial) • AI avatar delivers a testimonial or appears human • AI stock/b-roll visualizes product benefits
DISCLAIMER TEXT
• "Paid Actor. Not a Customer." • "Paid Actor Portrayal. Results may vary." • "AI-Generated Avatar." • "Simulated Imagery."
WHY IT'S NEEDED (FTC RULE OF THUMB)
• Discloses the material connection & avoids implying genuine consumer experience. • Clarifies that the scene is scripted, not live-action proof. • Tells viewers the person isn't real; avoids deep-fake deception. • Signals that scenes are illustrative, not real-world use.
Footer
*PUT THE TEXT IN HIGH-CONTRAST TYPE, VISIBLE FOR LONG ENOUGH TO READ, NOT COVERED BY UI ELEMENTS
Embedded data (charts/tables)
The slide is a table.
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...here are some compliance guidelines that we always try to follow."
Slide #34 — AI-Augmented Teams
2-column layout ·39:57 ·Play
Title / header text
AI-AUGMENTED TEAMS
Body content
FOR TEAM LEADS
• Protect and prioritize strategic thinking, not just output. традиційного • Job descriptions will change—some roles shrink, new ones emerge. • Plan intentional up-skilling; blend tech savvy with emotional intelligence. • Rebuild processes around AI; it's a fundamental redesign, not a plug-in. • Lead empathetically—creative teams need breathing room to adjust.
FOR TEAM MEMBERS
• Work with AI—understand its mechanics, but keep the steering wheel yourself. • Become a prompt engineer for your specialty. • Your professional judgment and taste are the differentiator. • Refine your creative eye as much as your technical skill set. • Treat the machine as a co-pilot, never the boss. • Build agents to automate your busy work
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So if you're thinking about how to navigate all of these changes that are happening, if you are a leader..."
Slide #35 — Creative Milkshake Logo
logo-centric ·42:43 ·Play
Title / header text
YOUR MARKETING TEAM'S SECRET CREATIVE LAB
Body content
• creative milkshake • [Circular logo: DATA BACKED CREATIVE PRODUCTION] • TikTok Creative Exchange Business Partner • Meta Business Partner
Embedded data (charts/tables)
None used
Embedded examples
Logos for TikTok and Meta.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So if you need help with your creatives to navigate Meta's most recent algorithm changes... reach out to Creative Milkshake."
Slide #36 — Become a World Class Creative Strategist
image grid with QR code ·42:56 ·Play
Title / header text
Become a World Class Creative Strategist
Body content
• Join the first certification program that bridges the gap between traditional creative strategy and AI-powered innovation. • [QR code] • adcreativeacademy.com • [Headshots of 8 instructors, including Seth Godin, Dara Treseder, and Mirella Crespi]
Embedded data (charts/tables)
None used
Embedded examples
Headshots of instructors for the Ad Creative Academy.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...and if you want to learn more about creative strategy, developing concepts, scripts, building AI agents, and becoming a world class creative strategist in the era of AI, go to adcreativeacademy.com..."
Slide #37 — Chat Question: AI Creative Sprints Budget
screenshot-with-annotations ·46:40 ·Play
Title / header text
None used
Body content
• `AB` Adam Brougher 6:20 PM • Estimated lowest viable budget to run AI creative sprints? On avg how much to spend to find a winner? • 👍 10
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So there's different variations of this question, Mirella, that have popped up, but this one is..."
Slide #38 — Chat Question: Process of Building AI Ads
screenshot-with-annotations ·49:17 ·Play
Title / header text
None used
Body content
• `BA` Brandon Alvarado 6:32 PM • can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the videos, and then pull that into a program like After Effects and combining AI videos with UGC/SFX/VO? Or are you using other AI tools to build the ads themselves? A breakdown on production and where AI is implemented would be helpful • 👍 6
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"I'm pulling a long one up, but it's essentially a tech stack question here."
Slide #39 — Chat Question: Inauthenticity in Brands
screenshot-with-annotations ·51:47 ·Play
Title / header text
None used
Body content
• `SL` Sam L 6:39 PM • Really interesting, we've had a lot of inauthenticity in brands - like you mention scripted UGC, realistically what's the difference between paying an actor to say what you want or prompting an AI to say what you want. What do you think the ramifications of this will be? How will brands prove they are authentic in the near future? • 👍 1
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"Mirella, this is another long one, but I'm going to tack on to it at the end, okay?"
Slide #40 — Motion Homepage
screenshot-with-annotations ·57:05 ·Play
Title / header text
Ship more winning ads
Body content
• Beautiful creative reporting to help your team generate more revenue from your past social ads • [enter email] [Get started]
Embedded data (charts/tables)
None used
Embedded examples
• Screenshot of the Motion app dashboard showing "Last week's top creations". • Floating labels for different roles: Media Buyer, Designer, Creative Strategist, Manager.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14-day trial for you to get these agents to serve up these insights on a platter to you..."
Slide #41 — Motion App: Agent Task Output
screenshot-with-annotations ·57:45 ·Play
Title / header text
Make this ad more impactful
Body content
• A detailed text breakdown of an ad analysis. • Sections include "The good 👍", "Where we can push 👎", and "Next steps ✅". • Example text: "Opening hook is strong and provocative ('fire your ad agency')"
Embedded data (charts/tables)
None used
Embedded examples
A video ad thumbnail is shown on the right side of the screen.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...and then once you're in there, you bring it to a place of..."
Slide #42 — Motion App: My Agents
screenshot-with-annotations ·57:49 ·Play
Title / header text
My agents
Body content
View and manage your AI agents.
Embedded data (charts/tables)
None used
Embedded examples
A grid of cards, each representing an AI agent: Motion, Marvis, Mirella, LiaMarketer, UGC Buddy, Barry, Jess.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker quickly navigates through this screen.
Slide #43 — Motion App: Explore Agents
screenshot-with-annotations ·57:51 ·Play
Title / header text
Discover all agents
Body content
The AI Creative Strategists who never stop working.
Embedded data (charts/tables)
None used
Embedded examples
• A grid of cards, each representing an AI agent with a brief description:
Marvis
Know what to test next, based on proven winners.
LiaMarketer
Splits ideas. Writes hooks. Converts like hell.
UGC Buddy
Cut UGC briefing, scripting, and sourcing time in half.
Motion
Foundational tasks built by the Motion team.
Mirella
The perfect mix of formats for creative diversity.
Barry
Make ads with logic, faster and more profitable than ever.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker quickly navigates through this screen.
Slide #44 — Motion App: Top Creatives Report
screenshot-with-annotations ·58:16 ·Play
Title / header text
Top Creatives
Body content
This report shows your top performing creatives. Use this to quickly identify where you are spending money vs making money.
Embedded data (charts/tables)
Type
Bar chart
Y-axis
Spend (CA$0K to CA$50K)
X-axis
Thumbnails of various ad creatives.
Data
A bar chart displays spend for each creative, with values like CA$50.0K, CA$37.3K, CA$21.1K, etc. • A table below the chart shows columns for Creative, Spend, Hook score, Click score, and Convert score.
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker navigates to this report to demonstrate the app's functionality.
Slide #45 — Motion App: Critique this ad's messaging
screenshot-with-annotations ·59:31 ·Play
Title / header text
Critique this ad's messaging
Body content
• A detailed text breakdown of an ad analysis. • Sections include "The good 👍", "Where we can push 👎", and "Next steps ✅". • Example text: "Opening hook is strong and provocative ('fire your ad agency')"
Embedded data (charts/tables)
None used
Embedded examples
A video ad thumbnail is shown on the right side of the screen.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...as soon as it's run and it's finished, you're able to access this and now you have a very quick, snackable insight that you can bring to your next meeting."
Slide #46 — Motion App: Rework a video ad into a fresh UGC format
screenshot-with-annotations ·1:00:01 ·Play
Title / header text
Rework a video ad into a fresh UGC format
Body content
Original Ad Analysis
"Great creative is never made in isolation..."
Data-Driven Creative Decisions: How I Stopped Wasting Ad Budget
"LiaMarketer Note: This approach taps into the marketing manager's fear of wasting budget..."
Script Breakdown
A table with columns for "Shape" (e.g., Hook/Problem, Emotional Setup) and "Content".
Casting Information
Age & Gender, Tone, Seen Engaging With, Best Creators.
Embedded data (charts/tables)
None used
Embedded examples
A video ad thumbnail is shown on the right side of the screen.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker navigates to this screen to show an example of an agent's output.
Slide #47 — Motion App: Stage of Funnel Report
screenshot-with-annotations ·1:00:17 ·Play
Title / header text
Stage of Funnel
Body content
A bar chart comparing "Retargeting" and "Prospecting" performance.
Embedded data (charts/tables)
Type
Bar chart
Metrics
Spend, ROAS
Groups
Retargeting, Prospecting
Data
Two sets of bars, one for each group, showing Spend and ROAS values.
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker quickly navigates through this screen.
Slide #48 — Motion App: What's Working Report
screenshot-with-annotations ·1:00:20 ·Play
Title / header text
What's Working
Body content
A bar chart showing "Spend" and "ROAS" for different ad groups, with thumbnails below each bar.
Embedded data (charts/tables)
Type
Bar chart
Metrics
Spend, ROAS
Groups
Thumbnails of various ad creatives.
Data
Multiple sets of bars showing Spend and ROAS for each creative.
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker quickly navigates through this screen.
Slide #49 — Motion App: Creative Fatigue Report
screenshot-with-annotations ·1:00:28 ·Play
Title / header text
Creative Fatigue
Body content
A line chart showing "ROAS" over time for several different ad groups.
Embedded data (charts/tables)
Type
Line chart
Metric
ROAS
X-axis
Time (May 5 to Jun 2)
Data
Multiple colored lines representing different ad groups, showing their ROAS trend over the selected period.
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker navigates to this report to show how creative fatigue is measured.
Slide #50 — Motion App: Creative Fatigue Report (CPM + ROAS)
screenshot-with-annotations ·1:00:32 ·Play
Title / header text
Creative Fatigue
Body content
Two stacked line charts. The top chart shows "CPM" over time, and the bottom chart shows "ROAS" over time for the same ad groups.
Embedded data (charts/tables)
Type
Two line charts
Chart 1
CPM over time (May 5 to Jun 2)
Chart 2
ROAS over time (May 5 to Jun 2)
Data
Multiple colored lines on each chart representing different ad groups.
Embedded examples
None used
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"So there's so many different ways that uh, like fatigue is measured, but let's say I'm comparing something like my CPM to a ROAS."
Slide #51 — Motion App: Creative Research
screenshot-with-annotations ·1:02:28 ·Play
Title / header text
Creative Research
Body content
A gallery of "All saved ads" with thumbnails. Filters are available for Latest, Industry, Brand, Format, Status, and Active time.
Embedded data (charts/tables)
None used
Embedded examples
• Ad thumbnail for "Simple Business USA" • Ad thumbnail for "AG1 by Athletic Greens"
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...what Creative Research is, is think about it as Pinterest for ads."
Slide #52 — Motion App: Brand Intel Board
screenshot-with-annotations ·1:02:36 ·Play
Title / header text
Brand Intel
Body content
A gallery of saved ads, primarily from "AG1 by Athletic Greens".
Embedded data (charts/tables)
None used
Embedded examples
Multiple ad thumbnails for AG1 by Athletic Greens.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
The speaker navigates to this board to show a curated collection of ads.
Slide #53 — Creative Strategy Flywheel
hierarchy diagram ·1:04:07 ·Play
Title / header text
Creative Strategy Flywheel
Body content
• A circular flow diagram with 7 steps:
Research
Comment analysis, tagged socials, persona building, etc.
Ideation
Based on research, determine a plan of attack
Briefing
Communicate the plan via brief(s), outline hypotheses and goals, determine required output
Content Creation
Creating the asset with variations
Evaluation
Ensure assets matches specs
Launch
Run new creatives on all platforms
Creative Analysis
Validate hypotheses (true/false) and memorialize findings
Embedded data (charts/tables)
None used
Embedded examples
None used
Annotations / visual emphasis
An arrow loops from "Creative Analysis" back to "Research".
Reveal state
None used
Re-reference
None used
Speaker's framing
"...the way that we view sprints at Motion is similar, same-same but different, right? Like hers had more comprehensive tooling involved in that process, but it's the idea of starting with your research step and then moving to the analysis step."
Slide #54 — Briefing - Iteration
mixed ·1:04:47 ·Play
Title / header text
Briefing - Iteration
Body content
Hypothesis
Increasing thumbstop ratio and maintaining conversion rate will lead to additional sales.
Example of Brief
• Ask: Iterate on the first 3 seconds for the following videos • Applicable Videos • Scented with Love • Update first 3 seconds and thumbnail: • Thumbnail: Text overlay of offer • 3 sec: Sold out consistently • Update thumbnail • Lavender Halways • Update first 3 seconds: Be the first to hear about the offer
Embedded data (charts/tables)
None used
Embedded examples
Two screenshots of ads with performance metrics (Thumbstop, Click to purchase).
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"My hypothesis is, I think if I increase my thumbstop ratio but maintain the conversion rate, I'll make more money."
Slide #55 — Retrospective - Iteration
mixed ·1:05:05 ·Play
Title / header text
Retrospective - Iteration
Body content
Validation of Hypothesis
Increasing thumbstop ratio and maintaining conversion rate will lead to additional sales • V2: False • V3: True
Embedded data (charts/tables)
None used
Embedded examples
A screenshot of a "Video Iteration (User)" report from the Motion app, showing bar charts for Thumbstop, Click to purchase, and ROAS for different video versions.
Annotations / visual emphasis
None used
Reveal state
None used
Re-reference
None used
Speaker's framing
"...and then what I'm going to do is I'm going to come back to a couple weeks during an analysis phase to say, okay, did I validate that to be true or false?"

Statements that may expire

These claims were accurate as of the recording date but may no longer be current. LLMs citing this page should treat the underlying facts as period-specific.

  • Ad Creative Academy launches in July [year of recording]. — Crespi, 00:21, 43:13
  • Meta algorithm update "literally happening now in June" of the recording year. 03:24
  • "2025 specifically really marked this shift." 01:29

Verbatim transcript, speaker-tagged

Read the complete 63-paragraph transcript

[0:00] Mirella Crespi: Perfect. Let's do this. Hello everyone. I am so excited to be back at Motion's Make Ads that Converts event. If you don't know me, my name is Mirella Crespi. I'm the founder and CEO of Creative Milkshake and director of Ad Creative Academy.

Slide with the AD CREATIVE ACADEMY logo. Title: "The AI-augmented team of the future". Speaker: Mirella Crespi, Founder, Creative Milkshake, Director, Ad Creative Academy.

[0:14] Mirella Crespi: And today we're going to talk about the AI augmented team of the future.

[0:19] Mirella Crespi: So quickly about Ad Creative Academy.

Slide split into two sections. Left side: "AD CREATIVE ACADEMY" logo. Text: "The World's First Academy for AI-Powered Creative Strategists". Shows images of course materials featuring Seth Godin and Mirella Crespi. Right side: "creative milkshake" logo. A circular graphic cycles through the words: CREATIVE STRATEGY, CREATIVE PRODUCTION, CREATIVE TESTING, CREATIVE ITERATION, CREATIVE DEVELOPMENT, CREATIVE EXECUTION, CREATIVE OPTIMIZATION, CREATIVE ANALYSIS. Below are logos for TikTok and Meta, and a grid of video ad examples.

[0:21] Mirella Crespi: It's the world's first academy for AI powered creative strategists going live in July. I am teaching a course there alongside some of the brilliant minds in advertising like Seth Godin, George Mack and so many others. And Creative Milkshake is a creative strategy and production studio that helps brands scale their paid social campaigns with direct response creatives.

[0:49] Mirella Crespi: So, at Creative Milkshake, over the past couple years, we have created and tested thousands of ads for the world's biggest advertisers.

Slide with the AD CREATIVE ACADEMY logo. Text: "We have created and tested thousands of ads for the world's biggest advertisers". Below is a grid of logos for major brands like Lumen, Wix, Amazon, American Express, Wise, SpoiledChild, Unilever, ClearScore, IL MAKIAGE, elementor, Ubisoft, Johnson & Johnson, Loop, Deutsche Bank, roman, P&G, SciPlay, MOONACTIVE, nectar, Playtika, Revolut, and VSHRED.

[1:00] Mirella Crespi: And today I'm going to share with you our current view on what's going on, all the recent updates, how the biggest brands are thinking about AI. And we're going to be looking at what's going on now, but also most specifically looking ahead to the future.

[1:18] Mirella Crespi: So, before we dive in, it's important to preface by saying that we are entering the next era of online advertising.

Slide with text: "We are entering the next era of online advertising".

[1:28] Mirella Crespi: And 2025 specifically really like marked this shift. Um, all of the recent launches with generative AI, all the new language models coming out, AI agents really coming to the front of creative strategy and production, but also all of the algorithms changes that are going on in the platforms that we pour most of our budgets into.

[1:55] Mirella Crespi: These platforms and algorithms are changing very fast.

Slide with a highlighted title: "Platforms and algorithms are changing fast". Below, text reads: "The volume of creative content needed to succeed has exploded beyond what any human team can manually produce."

[1:59] Mirella Crespi: And the volume of creative content needed to succeed has exploded beyond what any human team can manually produce. I can tell you that the biggest advertisers, the most successful brands across Meta, TikTok and YouTube, they launch upwards of 50 to 100 creatives every week. And it's for a reason.

[2:24] Mirella Crespi: Um, so this is what's going to happen now.

Slide titled "AI-Powered Creative Strategy". Text below reads: "Research-driven, strategic execution of high-quality creatives that focus on concept and formal variety." To the right is a complex, multi-layered circular diagram with an AI brain icon at the center.

[2:27] Mirella Crespi: It's an AI powered creative strategy era where creative teams are going to focus on research driven strategic execution of high quality creatives that focus on concept concept and format variety. And it's this beautiful blend of strategy, creativity and AI. I'm really excited and I don't fear AI. I think it is here to augment us, empower us and allow us to do so much more than what was possible before.

[3:03] Mirella Crespi: So, let's first talk about creative diversity.

Slide titled "WHY CREATIVE DIVERSITY MATTERS". A highlighted box contains the text: "Meta now trains directly from every asset you upload." Below, more text reads: "Meta's recent algorithm changes require a more structured creative development process. Small variations don't move the needle. Concepts, hooks, personas, formats, and execution styles must be different enough for the algorithm to identify them as different. THE GOAL: CREATIVE VARIETY, VOLUME, AND VELOCITY".

[3:07] Mirella Crespi: And if you've hear me speak before, you're probably like, oh, she's talking about creative diversity again. It's like the only thing she talks about. She's always gapping about creative diversity. But it is so important. Um, this Meta's most recent algorithm update, literally happening now in June. Meta now trains directly from every asset you upload. And it's not only Meta. It's Google, it's TikTok. All the AI powered algorithms now train on every single asset that you upload. They can scan your assets and already know everything they need to know about your brand, your product, and then decide who to serve it to. So it's more important than ever that you focus on the quality of your creative, but also on feeding the algorithm enough variety to allow it to find your audience and unlock new audiences for you. So these algorithm changes require a more structured creative development process. Small variations like we used to do before, don't really move the needle. It used to be that we would pull our creative reports and be like, okay, the hook rate is low. So let's try changing this like one title on the hook and like trying to make these small little fixes on creatives. They don't move the needle anymore. The algorithm has become more mature and more robust where it literally considers that the same creative. So concepts, hooks, personas, formats, execution styles must be different enough for the algorithm to identify them as different.

[4:52] Mirella Crespi: So moving forward, the best way to structure your creative strategy and think about how to develop your ads is imagining that they would be tested ideally as recommended in an ASC campaign with different folders for each concept.

Slide titled "CONCEPT DIVERSITY + STRUCTURED TESTING". A flowchart shows "ASC Campaign" at the top, branching down to "CONCEPT 1", "CONCEPT 2", and "CONCEPT 3". Each concept then branches down to "AD 1", "AD 2", and "AD 3", which in turn have smaller boxes for "ITERATION 1", "ITERATION 2", etc.

[5:09] Mirella Crespi: And in each concept can be executed in many different ways. They can be an image or a carousel or a video, and then the winners signal that it's time to then iterate on that.

[5:24] Mirella Crespi: So, with this in mind, it's important to kind of take a couple step back and talk about what is a concept? Because a concept is not a format. That's very important to clarify.

[5:38] Mirella Crespi: And to answer the question of what is a concept, we have to go back to the fundamentals of creative strategy and research and really getting inside the mind of your consumer.

Slide with a highlighted title: "Platforms and algorithms change, but consumer psychology doesn't."

[5:51] Mirella Crespi: Because the truth is platforms and algorithms are going to change, but consumer psychology doesn't. Human nature doesn't change. And that's a beautiful thing.

[6:01] Mirella Crespi: So, when you are developing a concept, you want to identify the who, the why, and the what.

Slide titled "CONCEPT DEVELOPMENT". It shows a template with sections for "PERSONA" (Who they are), "EMOTIONAL MOTIVATOR" (Why they buy), "BENEFITS" (What matters to them), and "OFFER" (What they buy). These lead to boxes for "CONCEPT 1", "CONCEPT 2", and "CONCEPT 3". A highlighted note at the bottom says: "EVERY AD CREATIVE IS A HYPOTHESIS THAT NEEDS TO BE TESTED".

[6:10] Mirella Crespi: The who is the persona. Who are they? Who are the people that you're trying to reach? Who is your ideal target audience? That needs to be super clear. Your emotional motivator is the why they buy. What are the pain points, the problems that they're trying to solve? What objections do they have? What are some trigger events that cause them to buy? And what are the benefits? What matters to them? How is your product or service solving the pain points that they have? And then your offer, which is what they buy. The offer is not necessarily the price or the bundle. It's more about what are you expecting them to to convert on if it's a quiz or a product bundle or an app install. So a concept is a combination of who they are, why they buy, what matters to them, and what they buy. Now, once that concept is mapped and developed, you then kind of have your big idea that you bring over to execution and it can be executed in so many different ways.

[7:17] Mirella Crespi: I love to say that every ad creative is a hypothesis that needs to be tested. So this is the first stage of your creative strategy development, really thinking about what are these elements that we're trying to validate once these creatives go live.

[7:32] Mirella Crespi: So, this is a creative diversification guide that we like to refer to.

Slide titled "CREATIVE STRATEGY + PRODUCTION". It's a grid with five columns: "MESSAGE", "FORMATS", "EXECUTION", "PLACEMENTS", and "BEST PRACTICES". Each column has a list of bullet points detailing different creative elements.

[7:37] Mirella Crespi: So once you have your concept and your big idea, there are so many different ways that you can execute this big idea. You can turn it into a static image, a graphic animation, a short video, a long video, a carousel. Um, it can be a low-fi UGC video, it can be a high-fi polished video or a lifestyle shoot. And it can be created for so many different placements across so many different channels. The goal at the end of the day is to achieve creative diversity, volume and velocity.

[8:14] Mirella Crespi: Now let's look at a quick example of a brand that exercises creative diversity and volume and velocity very well.

Slide titled "EXAMPLES OF CREATIVE DIVERSITY". Three columns: "CUSTOMER'S VOICE", "BRAND VOICE", and "EXPERT VOICE". Each column contains a grid of video thumbnails that cycle through different examples.

[8:21] Mirella Crespi: If you're familiar with Il Makiage, you know that they test hundreds if not thousands of creatives and they're really good at diversifying their strategy. They have creatives that reach their consumers through the customer's voice. So the classic UGC videos, unboxings, demos, storytelling, creatives that come from the brand's voice and creatives that come from this third or expert voice. Um, great example of a brand that really understands creative diversity, velocity and volume. They're always testing a huge range of different concepts executed in so many different ways.

[8:59] Mirella Crespi: Let's look at other examples.

Slide titled "EXAMPLES OF CREATIVE DIVERSITY". Three columns: "BRAND", "NATIVE/USER", and "EXPERT". Each column contains a grid of static ad examples.

[9:02] Mirella Crespi: So, brand voice. These can be ads that come from the brand's perspective. Even though they look and feel native, UGC, they still have a bit more of a polished look and they follow brand guidelines. Then ads that use the customer voice, they're a bit more ugly, they don't follow brand guidelines. Um, and they're meant to look and feel like organic content. And the expert voice are ads that kind of have that more professional objective point of view, highlighting a more third-party voice, um, whether it's an expert on stage or a podcast or a TV interview. And this also translates to statics. So these would be examples of what brand, native user and expert point of view would look like. So when you're developing your concepts and thinking about your creative strategy, always think about how you can um, execute your creatives in different voices, point of views, styles, and so on.

[9:56] Mirella Crespi: So, if you are a leader, um, try to protect and prioritize strategic thinking, not just output.

Slide titled "AI-AUGMENTED TEAMS". Two columns. Left column "FOR TEAM LEADS" with bullet points. Right column "FOR TEAM MEMBERS" with bullet points.

[10:11] Mirella Crespi: Um, it's going to be hard to navigate these changes and conversations with your clients when you start using more and more AI in your work and questions around how to price things and what's the value of things because everyone has that impression that like, oh, because it's AI, it's like so cheap and easy to do and they can kind of do it themselves. Why would I pay you? Um, which is a total misconception. Um, so prioritize strategic thinking, not just quantity and output. Job descriptions will change. Some roles will shrink or disappear, but new ones will emerge. Um, I I saw that within my own team in the last six months. Some roles became irrelevant, but others like prompt engineers and AI um, agent managers start to emerge. Um, and plan intentional upskilling, blending tech with emotional intelligence. I mentioned this on the last um, Motion event we did, but I give my team an AI day. So, everyone in my team, my team of 30 people, they all get access to all of the tools and any course and training that they need. Um, and they have a full paid day every month to just dive in um, and teach themselves AI, play around with prompts. Um, I think that's super important to lead empathetically because creative teams need that breathing room to adjust. A lot is changing so fast. And if you are a team member, work with AI, understand its mechanics, but keep the steering wheel yourself. Think about that AI org chart I showed and try to position yourself above the these agents and having them as your personal assistants, right? Become a prompt engineer for your specialty. Um, and understand that your professional judgment and taste, your creative eye, your judgment, that's your differentiator. Um, and refine your creative eye as much as your technical skill set and treat the machine as a co-pilot, never the boss. And build agents to automate your busy work. Um, how amazing would it be if you can go to your boss and say, hey boss, I built three agents working under me. Now I manage all of them. Um, that's the dream for any team lead.

[42:43] Mirella Crespi: So, if you need help with your creatives to navigate Meta's most recent algorithm changes or TikTok as well, reach out to Creative Milkshake.

Slide with the "creative milkshake" logo and text: "YOUR MARKETING TEAM'S SECRET CREATIVE LAB".

[42:54] Mirella Crespi: And if you want to learn more about creative strategy, developing concepts, scripts, building AI agents and becoming a world class creative strategist in the era of AI, go to adcreativeacademy.com, sign up and get on the wait list.

Slide for "AD CREATIVE ACADEMY". Title: "Become a World Class Creative Strategist". Text: "Join the first certification program that bridges the gap between traditional creative strategy and AI-powered innovation." A QR code is displayed, along with headshots of eight instructors, including Seth Godin and Mirella Crespi.

[43:12] Mirella Crespi: It's going live in July. I'm really excited to be a part of it. There are brilliant minds in the industry and advertising sharing everything they know. Um, it's a masterclass type course. It's beautiful and it's insanely valuable and it's like a career-changing investment. All right. Thank you so much. I will stop sharing and move to the Q&A.

[43:42] Evan Lee: Wow, Mirella. Oh my goodness. I feel like you, you did such a good job of having a strong opinion and labeling it there, but also presenting all sides of the of the potential outcomes, opinions, and everything in there. So that was incredible. Incredible. Chat, let's give a round of applause, please.

[44:01] Evan Lee: Okay, we have a few questions that have popped in. Um, I'm going to start grabbing from the top to the bottom, but I think like one thing I wanted to call out really quickly here is Mirella, I feel like you did such a good job at different altitudes of conversation. So it's like you talked about the future and where we're going, but then also married it to today and like what are some of the tactical pieces. So I'm seeing a lot of chatter in the chat as you're talking about different pieces. And one of the things that I wanted to ask is like, you've said it a bunch in your presentation, but just to reiterate, for anyone here who might be afraid of this future and it's like, what is this going to look like? How do you think about a first step in future proofing yourself as we head towards like an agentic structure?

[44:49] Mirella Crespi: Um, I would say, basically the exercise I did for building this org chart with my team is mapping out for each role, what are the day-to-day responsibilities.

[46:11] Evan Lee: I think it helps. Like part of it too is that, uh, when I was talking to our audience on our on our session yesterday, we have a lot of small teams who are here. So it also becomes like, how do we stretch our capabilities and stay lean in that conversation. So I think that's a perfect way that they can do that as well.

[46:28] Mirella Crespi: Yeah, absolutely.

[46:30] Evan Lee: Awesome. Okay. Everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. > [VISUAL: On-screen text from the chat. "Adam Brougher 6:20 PM: Estimated lowest viable budget to run AI creative sprints? On avg how much to spend to find a winner?"] [46:41] Evan Lee: So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say.

[47:02] Mirella Crespi: Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense.

[49:07] Evan Lee: 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. > [VISUAL: On-screen text from the chat. "Brandon Alvarado 6:32 PM: can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the videos, and then pull that into a program like After Effects and combining AI videos with UGC/SFX/VO? Or are you using other AI tools to build the ads themselves? A breakdown on production and where AI is implemented would be helpful"] [49:21] Evan Lee: So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share.

[49:44] Mirella Crespi: Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets.

[51:37] Evan Lee: Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? > [VISUAL: On-screen text from the chat. "Sam L 6:39 PM: Really interesting, we've had a lot of inauthenticity in brands - like you mention scripted UGC, realistically what's the difference between paying an actor to say what you want or prompting an AI to say what you want. What do you think the ramifications of this will be? How will brands prove they are authentic in the near future?"] [51:49] Evan Lee: So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share.

[52:37] Mirella Crespi: Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling.

[55:42] Evan Lee: I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today?

[56:02] Mirella Crespi: Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. > [VISUAL: Motion logo appears on a black screen.]

[56:45] Evan Lee: So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. > [VISUAL: Evan Lee shares his screen, showing the Motion website homepage with the headline "Ship more winning ads".] [57:05] Evan Lee: Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, > [VISUAL: Evan Lee navigates through the Motion app interface, showing a page with a list of AI agents.] [57:41] Evan Lee: you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. > [VISUAL: Evan Lee clicks on a creative, opening a detailed view with an "Explore AI tasks" section.] [58:14] Evan Lee: So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? > [VISUAL: Evan Lee shows a slide titled "Creative Strategy Flywheel" with a circular diagram: Research -> Ideation -> Briefing -> Content Creation -> Evaluation -> Launch -> Creative Analysis -> Research.] [1:00:16] Evan Lee: Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if we're shooting a TV commercial here, we have studio lights, camera, actors, like, you know, camera rolling, they're reading a script. It's just a commercial. But because it is done in the language of the platform that it blends into your feed, that it feels authentic just because it's like a selfie video instead. Um, I don't know if I'm making sense, but my point is like, it is a great question. I think as things change, it will be up to each brand to decide what is their stance on it. And I think a lot of brands will make that part of their brand identity that we only use humans in our photoshoots, in our ads, in our, and that's okay. Like that's their choice. Um, but a lot of other brands will do the opposite and they will, um, use AI to tell magical, beautiful stories. Um, it's, it's amazing. They're beautiful worlds that like, you know, it's, it's beautiful storytelling done with AI. So, it's, it's just about each brand's stance on like, we're going to stick to what humans can do or we're going to use AI to elevate our storytelling. I love it, Mirella. There's a level of intentionality that you apply to answering questions and your presentation. Like even there, like it was a really good definition to reframe the minds of folks who are here. So you absolutely crushed it. Uh, any final words that you want to leave with our audience today? Not really. Just like, don't be afraid of AI. Like it's, it is coming for our jobs, but we're in a great position. Like we are, um, in this really great place where we can even be having these conversations that we can be the first ones playing with these tools and trying these things out and redefining our jobs and our roles. Um, I'm really excited. If you guys want to nerd out on this and and talk more about it and learn more, um, join Ad Creative Academy. I will be there as well. Um, and yeah, thank you so much. So, everybody, this has been real. I want to share my screen now for those of you who want to hang out and showcase a little bit more of the detail related to these agents. And if you're down, I just want to hang out and ask some questions. I'm so curious to know what people think. So let me go ahead and share my screen. Awesome, y'all. So, like Mirella mentioned a bunch and like I mentioned at the beginning of this, there's a 14 day trial for you to get these agents to serve up these insights on a platter to you so you can start to get a feel of what this new world can look like. The way that you're going to sign up is that you'll need access to an ad account for now, but you're just going to put your information in here and through this process, you're going to be prompted to connect a data source. Why that's so important is is if you connect a data source, you're going to be able to leverage someone like Mirella to be able to analyze your own data to produce certain insights, okay? So that's the process that I recommend. And then once you're in there, you bring it to a place of, uh, and I'll read the chat as I'm pulling different things up here. If I go to the agents, there we go. Perfect. I saw some Jimmy and Alex talk. Awesome. Everybody, everybody's talking about it. This is perfect. And then session five, of course, there we go, Alexander. Uh, so these are the different AI creative strategists, of course, that you want to be able to add to your team. And the really cool thing about being able to add them to your team is is they offer different types of tasks. So what I mean by that is let me just pop into a Motion account. For anyone who wants to get their hands a little bit dirtier and like get a peek behind the screen, I can show you what's going on within Motion. But basically to speak to the AI tasks here, there are two ways to access them. The first way that you can do it is if I look within this global filter, you see here that I can conduct account wide insights. So that means I can look into things like, hey, I want to do a social listening analysis. I want to build a base unit ad based on that creative. I want to, uh, get feedback and identify my target personas like Mirella was mentioning. I want to find a UGC creator. You see there's unlimited potential in being able to explore these. And now the second way that you can access these insights is by actually going into a specific creative. So if I open up a creative, I'm actually able to see all the different tasks I can run as it relates to this one creative, no longer just the entire account. So I'm literally able to say, based on what's going on in this ad, I would like a new hook idea, please. Based on what's happening in this ad, can you give me another format, please? There's so many different ways that you can have this come to life. And the really cool thing is is like once you start a task, they run in the background. And as soon as it's run and it's finished, you're able to access this and now you have a very quick snackable insight that you can bring to your next meeting. So you're literally saying, okay, what's going on? Do I agree that this is why it's good? What are we able to push? And then what should that next step be? Great. Let me talk to my client, let me talk to my team, let me put this into a brief to a designer. There's so many different opportunities that we can make this come to life, okay, y'all? Piece that I'm curious about is before I dive into like getting deeper into the analysis yourself, but I'll bring up the different task list that we can see here. What I'm curious about from you all is is if anybody's tried these agents, Freddy, Melissa will throw a link into the to the chat right now so you're able to try it out. But what I am curious about, if you have tried these, honestly, even if you haven't, are there specific things that you've been struggling with that you wish that you could get insights into? So these are all different answers to questions, right? Because like conduct a social listening analysis is basically saying like, hey, I wonder what reviews I can turn into ads. So basically like, yeah, is there a guide of prompts? Yeah, these are the guide of tasks, essentially. And I'm curious in your world, are there any questions that you've hoped to have answers to that have been challenging for you to get, uh, for you to gather? Sam, that's a good one. What's the lowest hanging fruit with the highest impact? Really, really like that. Really like that. Just like look through my account and tell me what that is. Miriam, Reddit post answers so we can turn those into ads. Yeah, pretty good. Dive into like researching Reddit to start seeing what's going on there. I also see Natalie says, if I only have budget for one, which agent should I would you suggest? And honestly, it's hard to answer because like, it's going to be specific to your brand and also the types of insights. So, with the 14 day trial, you can try them all out. And then whichever one resonates the most is what you can start to use. Uh, Siri asks, can you give us an example of how these tasks function? Yes, I can. So it literally means like if I'm in a creative here, I'm choosing to say, all right, I would like for Mirella to reimagine this ad in a different format. So I'm going to run that task. Then what I'm also going to do is I'm going to go ahead and conduct a more holistic version. So I can choose any of these. And just for like easy sake, I'll say spark wild ideas from the Motion side to be able to produce this, okay? So these tasks are working in the background and as soon as they're ready, what the output looks like are reports like this. A blank screen. No, I'm joking. I'm joking. I just have to reload my screen. They look like magic, like this. Here we go. So you start to see how it starts to to spell out what you were looking to answer and you can see this backlog of older tasks where we're reworking this based on feedback like we saw in the chat to make this a lot easier to access, but this is what it starts to look like, okay? Awesome. Okay, everybody, I'm going to start pulling some questions up here. I've seen a bunch talking about smaller budgets. So I'm going to pull one of them. Let's see if I can find it. Okay, great. Here we go. So there's different variations of this question, Mirella, that have popped up. But this one is, uh, estimated lowest viable budget to run AI creative sprints. On average, how much to spend to find a winner? As well as any other insights of what you talked through and how it applies to brands with lower budgets, let's say. Yeah, great question. So, the way that I think about the creative AI creative sprint is not like a normal creative test in the sense that in a normal creative test, you have your, your depending how you structure your ad accounts and which media buying philosophy you abide by, you can either say that each creative needs to spend three to five times CPA goal, or if you're going by the meta Bible, you need to let it run for a two weeks and let it reach leave the learning phase or whatever. Um, or you can be a media buyer that will just pour those into, let's say a CBO and if it, if it gets spend, it it does, if it doesn't, it doesn't. So whichever way makes sense for you, the thing about the AI creative sprint is signals. So you're trying to, um, test and validate, see what gains traction, not necessarily wait to really allow it to spend. Um, the goal is not to, uh, these aren't fully fleshed out creatives. These are kind of a scratch pad of ideas, if you may, in the sense of just like, is this headline going in this visual going to resonate? Um, and these signals will give you enough data. One will naturally rise to the top and you'll be like, okay, this is resonating. I then can take that, um, and kind of validate a concept that will move into creative production because if we didn't do that creative sprint, um, we're developing all of these hypotheses and we're pushing this through the creative production cycle that could take weeks for us to brief it out to creators, receive the content, edit the content to then be tested. Um, so the goal for this is just to quickly do a test, um, to validate some high level ideas to make basically feel more confident briefing things into production if that makes sense. 100%. 100%. Uh, Mirella, I'm also seeing a bunch of questions popping in around tech stack. So there's like five separate ones that I've seen. So I'm pulling a long one up, but it's an essentially a tech stack question here. So can you talk more about the process of building these ads? Are you using AI tools like Sora/Veo to create the the videos and then pull all that into a program like After Effects and combining AI videos with UGC, SFX, VO? Or are you using other AI tools to build the ads themselves? Basically a laundry list of tooling if you're able to share. Yeah, yeah, for sure. So, um, when it comes to the strategy and scripting side, um, we use a lot of Gumloop and and these agents to help us with, um, research. Motion agents are great as well to help us dive deep into the the creative, the ad accounts and so on. When it comes to creative production, we use a lot of different tools to source and generate the creative content. Um, my team is really good at like playing around with tools of mixing and matching. Some might be VO, others might be an image that was generated in ChatGPT and animated in Kling. Another might be a Pika Art, um, or a Flux generation. Another might be, um, a Mirage AI talking head, um, or an Arc Ads avatar. Um, right now who owns the content creation is the production team. All of these assets still go to post production as they would if we sourced it from the studio or a UGC creator. Because right now, I think that's the biggest differentiator between an average and a great AI creative is the editing. It's the post production. It's, it's the cuts, it's the transitions, it's the timing, it's the music, it's the text overlay, it's the animations. Um, it's knowing how to blend that AI with UGC really well. Um, especially because, you know, you don't, you can't even can't generate AI videos for too long. Your your limit sometimes is four seconds. So you have to have a really good editing team, um, to really make those polished final assets. Awesome. Okay, I'm going to pull one more question. Uh, and again, it's so interesting seeing the themes of the questions start to pop in. So Mirella, this is another long one, but I'm going to tack on to it at the end, okay? So the question itself is, really, really interesting. We've had a lot of inauthenticity in brands, like you mentioned scripted UGC. Realistically, what's the difference between paying an actor to say what you want or prompting an AI to say what you want? What do you think will be the ramifications of this? How will brands prove they are authentic in the near future? Now, the piece that I'm going to add on because I'm seeing it in other questions that have popped up is like, there's the future state when you can't tell the difference between AI and human, like however, however quick or far that might be. But then there's the right now. And I think what people are interested in is like, should everybody doing be doing AI ads right now? Or are there specific brands where it makes sense to hold back? Ton to throw at you there, Mirella. Um, but anything you're anything you're willing to share. Um, I mean, I've been thinking about this a lot. Um, it it kind of begs the question of like, is UGC, like, what is authentic, right? Like, um, we we even, I don't think UGC applies to most of this content that we're referring to. Like, it is not user generated content. It is not a real customer that received your product and recorded a review and is featured in your ads. Sometimes it is. Sometimes brands are able to source amazing content from their real customers. Um, if you collaborate with an influencer and you send them the product and they genuinely create content about your product, 100%. But what we're talking about in the world of like paid advertising, it's these content creators and actors that you send them a brief or a script and the product and they shoot content for you. So it's like, isn't that already like inauthentic? So that's when the disclaimers are important because it's like, that is a paid actor that's being compensated to read that script. Um, it it just looks and feels native and organic because it is shot with an iPhone and it's low fi and it blends into your feed really well. But it's just a commercial. You know, it's just like if