How to Make AI Video Ads with Claude, Nano Banana & Veo | Will Sartorius, SelfMade

[0:00] Will Sartorius: [VISUAL: Intro animation for "Motion Presents" with a paper-cutout style landscape. Text on screen: "Motion Creative Strategy Bootcamp".] [0:03] Will Sartorius: [VISUAL: Title slide: "Building Your AI Creative Process", "Motion Creative Strategy Bootcamp", "Week 4". Speaker's name and title at the bottom: "WILL SARTORIUS, CEO, SELFMADE | Motion Creative Strategy Bootcamp 2028". A small video of the speaker, Will Sartorius, is in the top left corner.] So today we're going to be largely talking about AI creative, sort of shockingly. [0:07] Will Sartorius: Um, so, you know, first and foremost, who am I? [0:09] Will Sartorius: [VISUAL: Slide titled "Will Sartorius" with a photo of him and bullet points about his work: "CEO of SelfMade, a performance creative agency empowered by AI", "SelfStack newsletter, focusing on AI ad creative workflows", and "Building Skipper, a tool that generates on brand ads targeted to real buyer personas".] Uh, my name's Will Sartorius. Uh, I run a performance creative agency called SelfMade. Uh, we generate performance ads, shocker. Uh, and a lot of the ads that we're doing are being powered by AI. Um, I also have a newsletter that goes out every Friday. Uh, this is focused on AI workflows, uh, prompts, and things like that. Totally free. Uh, there is a link in the docs where you can, you know, subscribe to the newsletter and all that fun stuff. And then finally, uh, just because I'm a Renaissance man, jack of all trades, master of none, uh, I'm also building, uh, Skipper, which is a tool that does social listening, you know, collects your reviews, Facebook comments, subreddit comments, etc. Uh, synthesizes, synthesizes them into personas, angles, and emotions, and then subsequently generates ad creatives for you, uh, that you can, you know, in turn edit in platform. [1:05] Will Sartorius: [VISUAL: Slide with three columns. Column 1 is for "SelfStack" newsletter with a sign-up link. Column 2 is for "Socials" with links to LinkedIn and Twitter. Column 3 is for "Skipper" with a link to request beta access.] Uh, so those are sort of the three things here. Uh, you you'll get this deck afterwards, the SelfStack newsletter. Uh, my socials, I'm more active on LinkedIn than Twitter, uh, for, you know, probably pretty clear reasons. Um, and, uh, but if you want to just, you know, scan that QR code, get all that information, [1:20] Will Sartorius: [VISUAL: Slide with a large QR code. Title: "OR Just Scan This QR Code". A photo of Will Sartorius points at the QR code. A text box on the left says "Everything You Need Is Right IN HERE [LINK HERE]".] uh, you know, you can do that as well. Uh, also in the docs tab, uh, you can hit get all Will's stuff and you'll be able to, uh, access everything. Uh, cool. And all I ask is if you're ever in Brooklyn, uh, you buy me a beer, uh, for all this good stuff. So, pretty simple. [1:43] Will Sartorius: [VISUAL: Slide with text: "After This Call You'll Be Able To:"] Sweet. So my goal is after this call, you'll be able to walk away and do a few things. [1:49] Will Sartorius: [VISUAL: Slide titled "Clone Competitor's Ads". It shows two methods: "Quick Win Method" and "Scalable Method", each with before-and-after examples of ads.] You know, first and foremost, you know, clone a competitor's ad. And you'll see sort of throughout this presentation, we're going to be talking about two different methods. Uh, you know, the quick win method, uh, I'm not getting an award for that pros. And then we have, you know, the scalable method. The first half of this presentation, we're just going to talk about quick wins, things you can implement today to generate an ad. The second half of this presentation, we're going to sort of go into Claude code, go in more in depth and how you can create a repeatable system where you're generating, you know, 10, 20, 50 ads all in one setting. So, when I say clone competitors ads, what I really mean is repurpose. Cloning an ad is pretty ubiquitous and common vernacular within our space. What I really mean is taking an ad as inspiration and repurposing it to be your own. [2:42] Will Sartorius: [VISUAL: Slide titled "Animate Your Statics" with a "Quick Win Method" label. It shows two examples of static ads being turned into simple animated GIFs.] Uh, secondly, we're going to be talking about animating your statics. Uh, taking an existing ad and turning it into a GIF, you know, a generally an eight-second GIF. Uh, for those that are, you know, media buyers or what have you, uh, you know, CPMs on GIFs are generally lower than statics. So if you have a winning static ad, it makes a lot of sense to transform winning static ads into GIFs to get more leverage out of your, uh, ad creatives. [3:10] Will Sartorius: [VISUAL: Slide titled "Generate On Brand Static Ads (Scalable Method)". It shows four different static ads for "HEXCLAD" pans.] And then in the second half of the presentation, we're going to be talking about, uh, the more sort of scalable method. Uh, you know, building a system that has a foundation that you can repeat again and again and again. And so these four ads, I just asked Claude to generate four different types of static formats for me, and this was, you know, sort of one-shotted for each of these, you know, all for HexClad. Um, and so, you know, you can see the fonts are correct, uh, more or less, uh, and the product rendering is great. [3:42] Will Sartorius: So, we're going to get into the heady stuff again in the second half, and it's going to, you know, get a little heady, but, you know, if you can, try to stick with me. [3:44] Will Sartorius: [VISUAL: Slide titled "Multiply Ads Into Different Formats With AI" with a "Scalable Method" label. It shows one "JONES ROAD" ad being transformed into three different ad formats.] And then finally, we're going to talk about, you know, multiplying your ads into different formats. So if you have a winning static ad and you want to transform it into a, you know, from a headline ad into a before and after or into like a UGC style, you know, we're going to talk about that at the end as well. [4:06] Will Sartorius: [VISUAL: Slide with text: "Quick Win: A Lazy Man's Approach To Cloning Ads 😩".] Okay, cool. So let's sort of dive in. So quick win number one. Um, and I'm calling this a lazy man's approach to cloning ads. It's not that it's innately lazy. I think what a lot of folks get wrong about AI is that you can just sort of like prompt things and get things back. This requires some critical thinking. The best AI creatives you're going to get is going to come from a creative mind. So, I just want to, you know, lay that foundation. If you're able to be creative in this process, you'll be able to ultimately generate winning ads. [4:39] Will Sartorius: [VISUAL: Slide titled "Step 1: Upload the Ad to Clone to Claude (Opus Model)". Shows an ad for a wallet on the left and a detailed prompt on the right. The prompt has 9 numbered points for deconstructing the ad (AD FORMAT, COPY, PRODUCT, LAYOUT, etc.).] Okay, cool. So, step one, you need to find an ad you want to clone. So, you know, go into Motion, you know, find an ad creative, uh, from a competitor or for a similar industry. Then what you're going to do is you're going to take that ad and you're going to put it into Claude. I'm going to be talking about Claude a lot in this presentation. Um, you can do this in other LLMs, but my recommendation would be to do everything in Claude. Uh, at least right now, it's like by far the, you know, frontier model. Um, and everything we're doing with our agency, everything I'm doing with Skipper is largely done in Claude. So, what you do is you find a competitor ad and you upload it into Claude. And then you give it this prompt that I am very nicely handing to you. The idea is to deconstruct this ad, right? We want to pull out the ad format, the copy, the product, the layout, the background, the typography, you know, the different visual devices, the color palette, the spacing. What we aren't doing is just taking this ad and cloning it directly. We just want to understand the visual DNA of this ad. [5:52] Will Sartorius: [VISUAL: Slide titled "Step 2: Make It Yours". Shows an "Output" box on the left with the deconstructed ad elements, and an "Input" box on the right with a new prompt to recreate the ad for a different brand ("Ridge Wallet").] So we're going to get an output that looks somewhat like this, you know, a lot of M-dashes, of course. These outputs are going to be very long. Another thing that people get wrong is sort of like prompt engineering. Like the last time I read a full prompt was probably two years ago. I really don't read prompts in depth. And that's because often times what I'm doing is I'm taking my end result product, putting it back into Claude and having it refine the prompt, but we'll talk about this in more in depth. So you're going to see an output that is very, very long. Your next step within that same chat is to say, okay, you know, now I want to take what you just gave me and rework it for my brand. So my brand is Ridge. And very often what I would recommend here is put in the URL, um, as well as, excuse me, as well as a brand name. Um, and then a product description. I just sort of did like a barebones product description, but you can just go to the product page, copy the entire product page and put it there. The idea is we want to match all of these different elements with the proper elements of your product. [7:08] Will Sartorius: [VISUAL: Slide titled "Step 3: Gather Your Goods". Shows a flow: "Output From Step 2" (a text prompt) + "The Ad Clone" (image) + "2-3 of Your Ads" (images) + "Product image" (image).] Step three, sort of gather your goods. So now what we're doing is we're taking the output from this step, which is now a prompt for Nano Banana, and we're taking our ad that we're cloning, and then we're doing two additional things. So, how do we ensure that the AI gets our ads to be, you know, on brand? This, again, is the quick and dirty method. It works pretty well, um, but it's not necessarily scalable. So, what I recommend doing is go into your ad library, grab two to three of your own ads, ones that you think have very clear font hierarchies and your colors. So, you're effectively in this prompt telling, and you can see here, you're telling Nano Banana, and we can talk about Nano Banana in a second. I'm uploading two to three of my brand's existing ads. Match the exact colors, fonts, typography style, and visual tone from these. This is what my brand looks like. I'm also uploading a layout reference, you know, the ad clone, right? So, use this as a layout and the structure. You know, reference only, you know, positioning, spacing, composition. You're not going to have to do this. This prompt right here will generate the Nano Banana 2 prompt for you, right? This is a generalized template that will give you this working prompt. And again, this is going to be really long. It's going to be 800 to 1,000 words. That's okay. Prompts are meant to be long, and they're not meant to be read by a human being. [8:49] Will Sartorius: So, the idea here is we're going to be taking our prompt, we're going to be taking our ad clone that we're trying to mimic, we're going to be taking two to three of our own ads, and then we're going to be downloading the product image from the website. [9:04] Will Sartorius: [VISUAL: Slide titled "Step 4: Set Everything Up in Fal.AI". Shows the text prompt from the previous step on the left, and on the right, the Fal.AI interface with multiple images uploaded.] So, when all is said and done, this is what your setup will look like. You have the prompt in your in your input, and then you have your image URLs as well. So, you're uploading, in this case, I uploaded four. Uh, you know, you could also upload three if you had two references. Don't overdo it with the brand references because you may get some contradictory information. I would say two to three sort of max. Now, let's take a step back. I say, set everything up in fal.ai. You know, what the fuck is fal.ai? That sounds like really complicated, right? So, fal.ai is effectively a tool meant for developers, but honestly, I find it to be the very best tool to access every single model. You may have heard of like Higgsfield. Higgsfield, you know, I think it is a little shady. You know, I don't like working with them. You know, fal.ai is literally meant for developers, and what you can do in Fal is you can go pick your model and you can develop up to four ad creatives within it. If you don't want to do that, you know, I think it's 8 cents per generation in fal.ai. Um, you can go directly into the Gemini. Important caveat here. I am using Nano Banana 2. Before Nano Banana 2 came Nano Banana Pro, okay? Nano Banana Pro was very good at filling in holes and making assumptions about what you were trying to get at. Nano Banana 2 is not good at making assumptions. This is why you need a very, very long prompt. You can give these models so much context, and it will do an incredible job matching the context. So, don't, again, worry about giving too much information. When in doubt, overcommunicate. Something my sixth grade physics teacher used to always say, and I've sort of stolen it from from him. You know, I wonder how Mr. Abbott's doing today. Uh, but anyway, so the Nano Banana 2 prompts, very long, and you can see, I'm giving it a fuck ton of context here. Like, that is incredible, and it's matching really every aspect of this. So, Nano Banana 2 is your model. [11:10] Will Sartorius: [VISUAL: Slide titled "Step 5: Pick Your Favorite and Ignore The Rest". Shows the original ad on the left, and three AI-generated variations on the right, with thumbs up/down emojis.] And great, this was these were my first four generations, right? You can see like, not all of your outputs are going to be perfect, especially when we're sort of doing the sort of quick and dirty method of this. You know, I think this one probably turned out the best. And what you can see here is we reworked the copy based on the product description, right? We have the Ridge Wallet Pro. We have RFID blocking, we have track anywhere, we have quick card access, and a lifetime warranty. You can see these differ from the callouts in the original prompt. We are not cloning this verbatim. That would make no sense. We are repurposing an ad to be our own. And of course, you're going to get some like wonky things or like, you know, this like weird outline. There are ways to mitigate that, and we're we're going to sort of talk about that a little bit later. [11:58] Will Sartorius: [VISUAL: Slide with text: "Quick Quiz".] Okay, quick quiz. [12:01] Will Sartorius: [VISUAL: Quiz slide: "What's The Ideal Length of A Nano Banana 2 Prompt? 🍌" with multiple-choice options A-F.] What is the ideal length of a Nano Banana 2 prompt? Uh, and you know, A through F. So you can sort of put in the chat what you think is, you know, the right answer. I think I've sort of led the horse to water here, but, you know, you never know. Okay, let's look at our answers here. Okay, a lot of F's. Uh, some D's, some E's. Okay, so the correct answer is E. [12:44] Will Sartorius: [VISUAL: Answer slide: "Answer: 1000 - 1500 Words". Includes bullet points explaining why longer prompts are better for NB2.] You know, 1,000 to 1,500 words. Length actually does matter, especially with Nano Banana 2. You want to overload your prompt with context. It doesn't do a good job guessing. If you don't specify something, it won't do it. So the idea here is overload on context. Like if you want a single light stream touching the wallet, say that, right? If you just do a short prompt, it's going to fill in the holes in a really wonky way. If you want to just like use short prompts, use Nano Banana Pro, um, because that will make a lot of assumptions for you. But if you want to do this right, you need to really use Nano Banana 2. Uh, you know, there are there there is a bit of a ceiling here where you don't necessarily want to go above 2,000 words. Um, but again, like 1,000 to 1,500 words is I found is consistently the sweet spot. [13:41] Will Sartorius: [VISUAL: Slide with text: "Quick Win 2: A (Not So) Lazy Man's Approach To Animating Statics 😩".] Okay, cool. Quick win two. So, this is, you know, a not so lazy man's approach to animating statics. [13:50] Will Sartorius: [VISUAL: Slide titled "The Core Idea". Explains that you're not asking AI to animate, but to create two frames for AI to connect.] Because it actually isn't that lazy. Okay, so where I think a lot of folks get wrong, what a lot of folks get wrong here is, I'm going to upload my static into Kling or Veo 3.1 and I'm going to ask it to animate. That doesn't work. Like, what you need to do is you need to create two frames for your static ads. You have your static ad, and then you have the subsequent end frame, or the previous start frame. So, the first thing you need to decide is, is this ad that I have going to be the start or the end frame? Pretty simple. [14:24] Will Sartorius: [VISUAL: Slide titled "Step 1: Come Up With Ideas". Shows a static ad and a prompt to ask Claude for animation concepts.] So, here is a prompt to help you sort of get brainstorming. In this example, I actually knew what I wanted to do. I was scrolling through Jones Road's Facebook ad library. I saw this ad, and I was like, how funny would it be if like the little gummy bears sort of came up and and raised up the product. But if you are sort of having, you know, creative strategy block, if that's a thing, um, you what you can do is you can take this prompt, you can upload your image alongside the prompt, and you can ask it to give you a few ideas on how to generate a start or end frame. Again, I recommend coming up with a few ideas on your own just just because, you know, we're all human beings here. But like again, if you're really struggling, this is a good starting point. [15:13] Will Sartorius: [VISUAL: Slide titled "Step 2: The Nano Banana 2 Prompt For The Missing Frame". Shows a prompt to ask Claude to write a Nano Banana 2 prompt based on the chosen animation concept.] Now, it'll give you like three to five ideas of what to do. You know, use your product as a start frame and do this. You know, use your product as an end frame and do that. Pick one of the concepts that you like, and then use this prompt to generate the next step. So, you know, I want to animate this static ad. The concept is, and this is what I said earlier, you know, gummy bears march into the frame and lift the product up into position. My static is the end frame. Maybe a little redundant. My static is the end frame. So, now what I'm saying is, I need Nano Banana 2 to generate the start frame for my animation. So, again, I have the end frame, I need to generate a start frame. And of course, I'm going to use Nano Banana 2. It's the best image model out there. [16:04] Will Sartorius: [VISUAL: Slide titled "Step 3: Generate the Missing Frame". Shows the output from step 2 (a text prompt), the original "End Frame" image, and the newly generated "Start Frame" image.] And this prompt here is going to give me a Nano Banana ready prompt to utilize. So, all I'm going to do is I'm going to go into Nano Banana in fal.ai. I'm going to upload my prompt. I'm going to upload my end frame, and then I get my start frame. So, in this case, I know that what I want is I want the little gummy bears to sort of come in and lift the product. So, in this case, what the prompt gave me was, okay, so now we need to remove all the gummy bears, and we need to remove their respective shadows, and have the product be a little lower. So, now I have a really good starting point. [16:47] Will Sartorius: [VISUAL: Slide titled "Step 4: Get the Video Prompt". Shows the start and end frames being uploaded to Claude with a new prompt asking for a Veo 3.1 JSON prompt to create the animation.] Okay, back to Claude. So, we're switching between Fal AI, Claude. Fal AI, Claude. That's going to be a very consistent theme throughout this. If you want to get really cheeky, you can use Fal AI's API and do it all in Claude, but we can, you know, that'll be a conversation for for next month when you buy me a beer. Uh, so you're going back to Claude, you are pasting both the images that you just generated into the same chat you've been working on. And you're going to say, I'm uploading two frames for a video animation. This is my start frame. You have to be very explicit. This image is my start frame. This image is my end frame. The animation concept that I want to create, gummy bears march into the frame and lift the product up into its final position. Please generate a Veo 3.1 JSON prompt that creates a smooth 8-second animation from start to end frame. So, we're staying within Google's ecosystem here. We're just moving from static image to animated video. Again, you can do all of this in Gemini chat. You can swap between Nano Banana and Veo. [17:55] Will Sartorius: [VISUAL: Slide titled "Step 5: Generate Animation in Veo 3.1". Shows the prompt from the previous step, the start and end frames, and the Veo 3.1 interface in Google Labs.] But where I like to do this is directly in Google Labs. You can pay a very limited dollars to get unlimited, effectively, what is unlimited Veo credits. Uh, so, you know, this is what our team does. You know, our team is in Veo 3.1 all day long. Um, and so what you can see I'm doing here is I'm uploading my start frame as the start. It's going to ask you, do you want to upload a start and end frame? And you're going to say, yes. So, I'm going to upload my start frame, I'm going to upload my end frame. I'm going to say I want to do a video. I'm doing frames, like I just mentioned, rather than ingredients. Frames are better than ingredients with Veo. Just trust me on that one. I want 9x16, and I want to generate four. The other thing here, you'll notice this prompt is in brackets, it's prompt. What the fuck is this thing? Don't worry about it. Stop reading prompts. Stop trying to overengineer your prompts. It will drive you crazy, and it doesn't work. It does work, but it just there's much better ways to go about this. So, this, you'll notice, is a JSON prompt. The difference between Nano Banana 2 and Veo 3.1 is Nano Banana 2 does a really, really good job with text-based prompts. You don't need to do JSON prompts. Veo 3.1 is a little different. In order for it to adhere sort of frame by frame, you need to do JSON prompts. You can sort of equate Veo 3.1 to Nano Banana Pro, the version before Nano Banana 2, where it's making a lot of assumptions, and we don't want it to make a lot of assumptions. We want to tell it verbatim, this is what I want to have happen. And that's why you're giving it the JSON prompt. [19:38] Will Sartorius: [VISUAL: Slide titled "Step 6: Iterate (You Likely Won't One-Shot This)". Shows a "YIKES" result of the animation and a prompt for iterating by describing the issues to Claude.] And, you know, your first output may sort of suck. And like, that's okay. You know, this is all part of the process. So, like, unless you sort of one-shot this, which I can probably almost guarantee you won't unless you have a very simple animation, what you want to do is you want to take the GIF that you created. So, download the video as a GIF, you know, or download it as an MP4, convert it to a GIF, but no matter what, it needs to be a GIF. If you're if you're working directly in, you know, Google, uh, Flow, you can download it as a GIF directly. That's a great feature they have. You're going to put the ad that you created that was sort of bad into Claude, and you're going to tell it what the problems are with it. So, in this example, you know, we have this, it animates an additional green text. We have these gummy bears that look sort of like nothing like our original gummy bears. Uh, and then there's like this weird animation at the very end. Um, so, I just sort of outlined what the issues were here, and then I said, you know, give me a corrected Veo 3.1 JSON prompt that fixes these issues. And then maintain everything else that worked well, or things that I didn't mention. [20:50] Will Sartorius: [VISUAL: Slide titled "Step 7: Bask In Your GIF". Shows the before and after of the animated GIF, with the final version looking much better.] And eventually, you'll get there. This probably took me two or three tries, but then I got there, you know, and I was actually pretty happy with the end result. So, again, you're going to have to iterate a few times on this. Like, sort of similar to, uh, how Nano Banana Pro was. That's okay. Like, eventually, you're going to arrive at something you're happy with. Again, people assume that AI is just going to be immediate, like this is going to give me a great result. And when it comes to statics, yes, we are there. When it comes to videos and GIFs, we're getting there. Veo 4 comes out in May. This presentation will be obsolete at that point because we're going to be behind. Um, but at the end of the day, uh, just be patient. Just be patient. And like, an animating an aesthetic like this, you're you are effectively not guaranteeing, but shouldn't use that word. You are taking a winning ad and you're converting it into a new format, so it's going to hit a new audience, uh, with presumably lower CPMs. [21:53] Will Sartorius: [VISUAL: Slide with text: "Quick Quiz". The image of the speaker is now upside down and giving a thumbs down.] Okay, another quick quiz. [21:56] Will Sartorius: [VISUAL: Quiz slide: "When Do You Need to Use JSON Prompts?" with multiple-choice options A-E.] Uh, when do you need to use JSON prompts? Nano Banana 2, Veo 3.1, Claude, Kling, or all of the above. Sweet. I think most of you are getting this right. [22:34] Will Sartorius: The answer is B, Veo 3.1. [22:37] Will Sartorius: [VISUAL: Answer slide: "Answer: Veo 3.1". Includes bullet points explaining that while other models accept plain text, Veo 3.1 requires structured JSON for precise control.] Uh, Nano Banana 2, as I've mentioned, is very, very text-based oriented. You can use JSON prompts. It's not going to give you, um, it's not going to really help your output that much. Uh, Claude, obviously, you're not going to be using JSON prompts. Kling, also very good at text-based. We're not going to talk about Kling today. We're just going to, you know, sort of keep things simple. Um, the world we're working in now, Veo 3.1 is really the only one that requires a JSON prompt. You can still use text-based prompts in Veo 3.1. Absolutely, absolutely. But you're just going to have less control over the output. It's going to make a lot of assumptions, and we don't want these models making assumptions. We want to be in control. [23:26] Will Sartorius: [VISUAL: Slide with text: "Time to Get Heady 🍻🍻", "Using Claude Code to Build Iterative Ad Creation Systems", "The Scalable Method".] Okay. So, that's sort of like the easy stuff. Now it's time to sort of get heady. So, now we I want to talk about like the the scalable system here. So, we talked about quick wins in animating statics, in cloning competitor ads. Now I want to talk about getting into Claude code and building like iterative ad creative systems. So, we can build a foundation to ultimately generate as many statics as we want, uh, you know, as many animations as we want. This setup, while it may seem complicated, is really not that bad. [24:10] Will Sartorius: [VISUAL: Slide with text: "Please Realize, People Try to Overcomplicate This 💩:", followed by "Prompts = Just Text", "Markdown Files = Just Text", "Agents = Just Text", "Skills = Just Text".] Like, I think what people need to realize is people love to overcomplicate things to make themselves seem smarter than they are. This is really not that scary. Like, prompts are just text files. Everyone's familiar with a prompt. A markdown file is just a text file. An agent is just a text file. A skill is just a text file. These things are not that scary. Again, people will say, oh, you, you know, I want to share that skill, I want to share that markdown file. All that really means is they're sharing a something you can open and text edit, right? So, don't be scared when people say skills, agents, markdown files, like these are all just tools to instruct Claude to do a certain thing. Let me add, you know, one or two more things here. Yes, Claude can do a lot of things on its own, right? You can say, analyze this brand and give me its personas. The point of having like skills or markdown files is to sort of catalyze that effort and honestly use fewer tokens. If you can direct Claude directly to a skill or a markdown file that outlines this is how you do something. You know, this is how you open a Microsoft Word document. This is how you generate a brief. This is what a headline ad looks like. It's going to work much faster, it's going to use fewer tokens, and then you're going to have a file that you can iterate on over time. [25:42] Will Sartorius: [VISUAL: Slide with text: "So Don't Panic! I'm Giving You (Mostly) Everything You Need 👉", with the same QR code as before.] So, don't panic. I know we're getting heady, you know, love Heady Topper, Focal Banger's also great. I'm again, giving you everything. This is the same QR code as earlier. I'm giving you all these prompts. I'm giving you this system. So, if this seems scary or, you know, you're overwhelmed already, that's fine. You know, we're going to be working through this together. [26:04] Will Sartorius: [VISUAL: Slide titled "Proper Prompting Requires A Proper Setup". Shows four boxes: "Brand Extraction Prompt", "Brand Reference Cards", "Format Templates", "(Optional) Copywriting Agents".] So, proper prompting requires a proper setup. So, this is a workflow I found when Nano Banana 2 came out, and it works really, really well, and I've just been continually adding to it. Even since I did this slide, uh, I've I've another step I like to add, right? But sort of the four core things you want to have in your Claude project. One, a brand extraction prompt. Uh, you know, people over the internet are talking about this brand extraction prompt. I can almost guarantee sort of came from my newsletter. The idea here is what we're doing is we're taking a prompt, we are telling it to extract as much information as humanly available from a specific brand, and we are saving that in a folder. We're going to go we're going to go in depth into each of these things, so so don't worry if what if nothing I'm making saying makes sense right now. Uh, brand reference cards. Sort of similar to when we were cloning the ad and we uploaded references for our own brand. A much more efficient way to do that is to upload a brand reference card that has all of your fonts, has all of your colors, uh, it has like how you like your products to look, and you upload that as an image rather than uploading that within the prompt itself. Format templates. This will help you immensely. Again, Claude can go do this research for you, or you can just give it, give it to it up front. So, what a format template is is it says, this is what a headline ad looks like. So, when I'm generating a brief, I'm going to go look at that format template for the headline ad, and I'm going to match the copy style to the headline structure. So, the idea is we are telling Claude exactly how to brief an ad before it briefs it. Step four, you know, after the brief is done, you can spin up agents. I cannot tell you how easy it is to spin up agents. You literally go into Claude and you say, can you generate five agents for me that do the following? And you say, agent one is focused on persona, maybe. Agent two is focused on emotion, agent three is focused on angle, agent four is, you know, uh, using Evan Lee's voice and ensuring that we're getting it properly, right? You know, an agent five is about grammar. What it will do is it will take your copy, it will run it through these agents, and it will score it until you get a proper piece of copy that actually works. So, that's the foundation, right? Now, every time, this is what you want to, this is what you want to run anytime you sort of go back to that project if you're in chat, or you go back into that folder when you're in code, you pick your format. So, this is where the you're going to say, I want to create a headline ad, if that's one of your templates. And also, you know, I'm giving you the headline template. This is going to be your starting point. So, you can upload your headline template into Claude code, and you can say, I want to create an ad with the headline template focused on this persona, this angle, and this emotion. This part of the process will only work if this component is done. Then what happens? Claude will write the brief for you. It will put in copy, creative direction, uh, all from that template. So, it'll match the copy format based on the template. Then you have your agents or your assistant score it. Uh, again, very simple prompt, just say, can you please have my agents review this copy? Once you get copy, you're happy about, you ask it to turn it into a Nano Banana 2 prompt. Pretty easy. Again, when I was putting this presentation together, I built this whole system for HexClad, and this was the the first ad I got for uh, for the headline. You know, one pan, every technique. See you like cast iron, release like release like non-stick, outlast both styles. Like, that's really fucking good copy. [39:12] Will Sartorius: [VISUAL: Slide with text: "Layer 1: Brand Extraction".] Okay, so let's talk about the sort of layer one brand extraction. [39:14] Will Sartorius: [VISUAL: Slide titled "Layer 1: Brand Extraction Prompt". Left side explains what it is, what it outputs, and why it matters. Right side shows a screenshot of a very long, complex prompt.] Again, I'm giving you all of this. I'm giving you the brand extraction prompt. You know, this is a lot of sauce. Um, again, we've talked about what this is, but effectively, this is just going and reverse engineering what a brand is, what they stand for, uh, you know, what their products are like, right? It's just building that initial foundation for the brand. [39:40] Will Sartorius: [VISUAL: Slide titled "Note: Claude Sucks At Extracting Proper Fonts + Colors". Lists two options: 1) Inspect the webpage, 2) Upload brand guidelines.] This is an important note. Claude really sucks at extracting proper fonts and colors. So, when you're running this brand extraction prompt, if you already have brand guidelines, put that into the project if you're working in chat, or put that into the folder if you're working in code, right? That will ensure that when we're generating the spec cards, we're we have the proper fonts and colors. So, that's option two. [40:10] Will Sartorius: [VISUAL: Slide titled "Layer 1: Brand Extraction Prompt (Getting Fonts + Colors)". Explains the process of inspecting a webpage (Jones Road) to find font and color information.] Option one is if you don't have the brand guidelines and maybe you want to do some spec ads for a client and you really want to impress them. This is like the sort of the hack I found is you go to the client's website, you inspect the web page, you type in font up here, you copy and paste all of that, who knows what it means? You know, I'm not an HTML coder. You put that into Claude and you say, update the spec cards accordingly, right? And you just repeat that process with color. [40:38] Will Sartorius: [VISUAL: Slide titled "Note: Save The Extraction In Your Project / Folder". Shows a meme from Zoolander ("The .md file is IN the computer"). Explains the difference between saving the file in Claude Chat vs. Claude Code.] Okay, so this is going to be the first of many reminders. When it does that extraction, if you are not working in Claude code and you're working in Claude chat, it's going to present a file to you. Save that file and put it in your project. This is essential. If you forget this step, nothing else is going to work, right? If you're working in Claude code, this will automatically happen. If you're working in Claude chat, it will give you a file, save that in your project. [41:08] Will Sartorius: [VISUAL: Slide with text: "Layer 2: Brand Reference Cards".] Okay, brand reference cards. [41:10] Will Sartorius: We've talked about these a lot, but they're very important. [41:12] Will Sartorius: [VISUAL: Slide titled "Layer 2: Brand Reference Cards". Compares "What Most People Do" (Quick Win Method) with "What Actually Works" (Scalable Method).] So, you know, what most people do? We did this at the beginning. We just uploaded reference images to the model, sort of like fingers crossed, I really hope this matches the aesthetic. Um, but we're not giving it enough information, right? We're we probably have like a 60, 70% hit rate with the quick win method. When we're doing the scalable method here, my hit rate is probably between 90 and 95%. So, that's pretty good. What I'm doing is I'm uploading those brand spec cards and the visual style cards, and I'm feeding these into Nano Banana 2's images. These include the hex codes, the type, the the, you know, photography style, etc. [41:56] Will Sartorius: [VISUAL: Slide titled "Layer 2: Brand Reference Card Prompts". Shows two prompts: "Prompt for Brand Spec Card" and "Prompt for Visual Style Card".] So, again, this is just the brand reference card prompts. This is how you generate the spec card, this is how you generate the visual style card. I don't think I need to spend too much time on this. [42:08] Will Sartorius: [VISUAL: Slide titled "Layer 2: Brand Reference Card Examples". Shows the output of the previous prompts: a "Brand Spec Card" and a "Visual Style Card" for Hexclad.] Um, the one caveat is it will, uh, give it to you as an HTML, but very kindly, I asked it to convert it to a PNG at the end. Again, these are what the cards look like. So, we did this for HexClad. You can see we have the logos here. We have all of the proper fonts, like headline, pretty small. The UI, the subheaders, the body copy. We have all the colors, we have the do's and don'ts, we have what the CTA's look like. So, this is the brand spec card. And then on the visual style card, we have like what the photography direction should be. We have styling rules, we have a founder quote, we have example ad breakdowns, like what an ad should look like. And then we have like visuals, do's and don'ts. So, if the sort of TLDR is, the spec card is telling you what your brand looks like. The style card is telling you how your brand feels. What's this brand, you know, what's this ethos? [43:07] Will Sartorius: [VISUAL: Slide titled "Note: Save The Cards In Your Project / Folder". Shows a cartoon scientist and explains the importance of saving the generated cards.] Again, save the cards in your project and folder. It's really fucking important. If you don't do this, nothing that we're going to be doing next matters. If you're in Claude chat, again, download the file that it created for you and put it in your project. If you're working in Claude code, it will save automatically. [43:23] Will Sartorius: [VISUAL: Slide with text: "Layer 3: Format Templates".] Okay. [43:25] Will Sartorius: [VISUAL: Slide titled "Layer 3: The Format Template". Explains the concept of a format template as a "recipe card" and shows an example of a "Static Headline Template".] Finally, oh, sorry. Step three, format templates. So, think of this as a recipe card. If someone hands you a cookbook, you're going to have a much better sense of what to cook. But if someone said, you know, make, uh, I don't know, chicken teriyaki for me, you're going to be like, I don't know where to start. So, effectively, what we're doing is we're giving Claude a cookbook on how to generate all of these different ads. And within this cookbook, we have all of these different recipe cards. These are the format templates. I, the way I've gone about this is I've created format templates for different ad types. There's a lot of different things you could probably do here. I've thought about doing one for hooks, like, you know, maybe a hook style. I've thought about doing one for winning ads. Uh, there are so many different possibilities. But again, you just like want to have a cookbook with so many different recipe cards where you are saying, this is what this style looks like, this is what that style looks like, etc. And again, I'm giving you a headline template to get you sort of started on this. [44:31] Will Sartorius: [VISUAL: Slide titled "Layer 3: How to Build Your Own Format Template". A 3-step process: 1) Collect Examples, 2) Upload to Claude, 3) Refine the Output. Shows examples of headline ads.] How would you even go about building your own format template? This sounds super intimidating and super daunting. The way I go about this is I collect a lot of ads that are utilizing a similar format. So, these are all headline ads. Here are six headline ads here, right? I think I collected around 5,500 ads, and, you know, I just used this tool, uh,

[46:02] Will Sartorius: [VISUAL: Slide titled "Note: Stop. Save It. (Yes, Again.)". It shows a cartoon soldier yelling "I SAID SAVE IT." and the text "You thought I was kidding the first time?". Below are instructions for Claude Chat and Claude Code.] Again, save it. Take what you created, whether it's a headline template, you know, whether it's uh, you know, a hook template, doesn't matter. You're building a cookbook. You need to have a lot of different templates in here. You are just making Claude's life a little bit easier so it doesn't have to guess, right? You are in control. That is like the fundamental thing I'm trying to teach here is like, you are in control of what this model is doing. It is just helping you get from point A to point Z. Again, this is why critical thinking throughout all of this is essential. We can't just assume these models are going to do it for us, but if you follow these steps that I'm outlining now, you will create a repeatable process that will spit out ads that you will be mind-blown by. [46:52] Will Sartorius: [VISUAL: Slide with a light green background and blue text: "Layer 4: (Optional) Copywriting Agents"] Okay, layer four, the copywriting agents. [46:55] Will Sartorius: [VISUAL: Slide titled "Layer 4: (Optional) The Copy Scoring Agents". On the left is a blue box titled "Think of it as a copywriting QCer". On the right is a box titled "This Is What One Looks Like" showing a document example for "Agent 01: Persona Fit Assessment".] Think of this as just like a QCer, right? This is just a text file that tells you to review the copy. So, in this case, I like, you know, a persona fit is my first agent. Like I so, you know, what's the role? You're a consumer psychologist specializing in persona-based ad targeting. Your job is to assess whether the ad copy resonates with a specific persona, not whether it's, you know, good copy, whether it speaks to the language, addresses the pain, and matches the worldwide view of the exact human it's targeting. You'll be given a persona definition and an ad brief. Your job is to score how well the brief speaks to that persona and provide actionable fixes to close any gaps. So, your idea here is like you want to build an army of agents or assistants or QCers, whatever you want to call it, that you can call on every time you write a brief. So it will do a much better job generating actionable copy that is persona driven, that has the right angle in mind, uh, that has the right emotion in mind, um, and maybe is on brand. [48:01] Will Sartorius: [VISUAL: Slide titled "Layer 4: (Optional) The Copy Scoring Agents: DIY". On the left is a bulleted list of focuses. On the right is a box titled "Prompt for Building An Agent" with a prompt template.] So, how would you do this? Again, there are so many different ways you could go about it. It's you could write one sentence into Claude code and say, hey, I want to generate an agent that is focused on brand voice. Uh, can you spit that out for me? I that's generally how I started doing this and just sort of playing around with this, but now that I've, you know, spent a lot of more time in here, this is what where I would sort of start. Like, I need a copywriting assistant that reviews my ad copy for, you know, whatever it is your focus. Sorry, this also should be highlighted. Oh no, no, sorry, that doesn't matter. Once you once you fill in this, the rest will be uh, go through. And the idea is when you have all these agents together, right? They will continue iterating on your copy until you have something that passes each different level. [48:54] Will Sartorius: [VISUAL: Slide titled "Layer 4: (Optional) The Copy Scoring Agents: My Favs". It shows seven agent types (Persona, Angle, Emotional, Brand, Conversion, Format, Grammar) and a flow chart for the scoring process.] So, this may take some time. Um, you know, the for for, you know, a bureaucracy of agents to score, you know, an ad brief, 15 minutes, maybe 30 minutes, I don't know, depends how complicated the ad brief is. You could create agents for reviewing UGC scripts. You could create agents for, uh, you know, sounding, I don't know, like you're from New York, right? There are so many different ways to sort of go about this. These are my sort of personal favorites. Uh, they are ever expanding. I think I have like 20 copywriting agents now. Uh, and they all have beef with one another. Um, but the idea here is, you know, persona, angle, emotion, is it fit the brand? If you do a brand fit agent, the cool part is, if you're working in Claude code, it will know from the previous steps what the brand is all about because we did the brand extraction, we generated the spec cards, right? This is a process that we are building upwards. Um, conversion agent, you know, is this ad going to is this copy is this copy conversion focused? Is it formatted properly? You know, is it is it, you know, does it have proper grammar? Um, these things are all super, super important. Uh, and my grammar agent, I ensure there are no M dashes. Um, you know, RIP. But, you know, I think maybe it'll come back in droves at one point in our lives. But, the idea here is the brief will enter the the scoring, the agents will all review, and it'll keep iterating until each one gets sort of like a 90 plus. [50:41] Will Sartorius: [VISUAL: Slide titled "Note: Stop. Save It. (Last Time, I Promise.)". It shows Goku from Dragon Ball Z and the text "Four for four. You're building a system." Below are instructions for Claude Chat and Claude Code.] Okay, this is the last time I'm going to remind you. Uh, save it. You know, we're four for four, you're building a system. If you're in Claude chat, download it, put it in your project. If you're in Claude code, it will already save to your folder every time you create an agent. [50:57] Will Sartorius: [VISUAL: Slide with a light green background and blue text: "Are You Still With Me? 🥺 ... Now What?"] Okay, exhale. We made it through the heavy part. Hopefully most of you are still with me. Um, so now what? What happens next? [51:10] Will Sartorius: [VISUAL: Slide titled "Generate On Brand Static Ads (We Back)". It shows four example ads for Hexclad pans.] Okay, so do you guys remember this slide from the beginning? So now we want to be able to generate on brand statics. We are back, baby. So, I think what I did here was I had a template for four different ad types. I said, generate briefs for this persona, angle, and emotion. I want a headline ad, I want a uh, statistical ad, I want a testimonial, and I want a features and call outs ad. So what it did was it went, it reviewed all of my, it went into the cookbook, it said, what does a headline ad look like? You know, what does a testimonial ad look like, etc. It generated a a brief for based on that cookbook. It called in the agents, it said, review the copy, then gave me a brief. And if I'm cool with that brief, I'm just going to take that brief, convert it to a Nano Banana prompt, and then I'm going to toss it into Fal. [52:08] Will Sartorius: [VISUAL: Slide with a light green background and blue text: "Let's Do It For JONES ROAD 🧴"] So let's just do this for Jones Road. We're trying to we're trying to do all three of the main motion brands here. [52:13] Will Sartorius: [VISUAL: Slide titled "Let's Start Briefing!". It shows a "Starter Prompt for Briefing" on the left and the resulting "COPY" and "CREATIVE DIRECTION" on the right.] Okay, so let's start briefing. So this is sort of where everything comes together. Here's a sort of starting point for your brief. This is assuming you've done a good job saving all of your previous information into your project if you're using Claude chat. If you're using Claude code, presumably it's already all in there. You know, I want to create a headline ad, a headline is one of my templates, one of my recipes for the Jones Road shower gel. That's what I have here. The persona I want to focus on is a sensitive skin sufferer. My angle is transformation, and the emotion I want to uh, exhibit is relief. Use my brand bible and format template to write the full brief: headline, subhead, copy, creative direction, everything. Follow the format template directly. So, great. So now it's going to it's going to go, it's going to think, it's going to give me some copy back. Your skin barrier has been begging for this. Dermatologist tested, 2,700 plus harsh chemicals eliminated. Lathers rich, rinses clean, never strips. It's a lot of copy for a headline ad. Um, and then it gives us creative direction, like what should the ad look like? So, generally speaking, the first batch of copy I get back, I don't even read. I I'm like, this needs to go to, you know, my my good friends, you know, the agents. [53:39] Will Sartorius: [VISUAL: Slide titled "Let's Start Briefing! Tag In The Agents 🕵️‍♂️". It shows a flow from a "Starter Prompt for Agent Review" to "Final approved copy (v3)" and then to a "Starter Prompt To Convert It To A Nano Banana 2 Prompt".] And so, then what I say is, please have the agents review the copy. Please iterate on the copy until every agent gives it a 90 plus out of 100. So, this is going to, you know, for my SpongeBob fans in here, this is going to take some time, you know, 20 minutes, uh, on average, maybe less, maybe less depending. Um, and, you know, go do other things, you know, get a cup of coffee, pet your dog, whatever. Um, eventually you'll get sort of like a final approved copy. You're going to have a headline, you know, it's not your sensitive skin, it's your shower gel. Ooh, like that sounds super compelling to me. Like, maybe I I was like maybe super itchy. Maybe my shower gel is the issue. I didn't even consider that as an option. You know, the headline, 2,700 plus irritants eliminated. What's left lathers rich, rinses clean, and finally doesn't sting. Like, that is talking to my sensitive skin persona. It's talking about sort of a transformation and it's providing the the emotion of relief. Great. So now I have this final approved copy. It's given it to me. The last step is take the brief that we just created with the copy and convert it into a Nano Banana 2 prompt. Again, I'm giving you a prompt to generate a Nano Banana 2 prompt. Um, you know, take a drink every time I say prompt. So, convert this into a ready to paste Nano Banana 2 image generation prompt. Uh, this, I'm assuming we're all going to sort of start with 9 by 16. You can convert a Nano Banana 2 from 9 by 16 to 1 by 1 or 4 by 5 very, very easily. Um, again, I'm not I'm not going to read all this, but the idea here is you're taking the brief that you just created, right? You're converting it into a Nano Banana 2 prompt. You're taking your Nano Banana 2 prompt, and then you're uploading it into Nano Banana 2. [55:32] Will Sartorius: [VISUAL: Slide titled "Let's Start Briefing! Upload The Prompt, Product, Spec Cards And Generate! ✨". It shows screenshots of the prompt and image upload interface for an AI tool.] So this is what this is going to look like, right? Here's the prompt that we just generated. Here are our two spec cards, right? Uh, that we have with the colors, the fonts. Um, here's brand spec, here's visual spec, and here's the product. [55:47] Will Sartorius: [VISUAL: Slide titled "Successfully One-Shotted 🤯💥". It shows two generated ads for Jones Road shower gel, one in 9:16 and one in 1:1 aspect ratio.] And yeah, that was my first ad that I generated. Uh, it was, you know, the copy is pretty on point. The product looks great. Um, give sensitive skin a break. Like this matches pretty close, if not almost identically to sort of Jones Road. Once I have my 9 by 16, I just change Fal AI from 9 by 16 to 1 by 1 and I ask it to resize it. [56:12] Will Sartorius: [VISUAL: Slide with a light green background and blue text: "Last Thing 🙌: Let's Multiple This Concept Across Formats"] Okay, last thing, and then we can chat. Let's sort of multiply this concept across multiple formats. [56:20] Will Sartorius: [VISUAL: Slide titled "Multiply Ads Into Different Formats With AI". It shows one ad on the left, an arrow, and three different ad formats on the right. A text box overlays the middle saying "Remember this Slide?".] So, remember this slide at the very beginning, about like an hour and change ago? We had an ad, right, that we like. Let's assume we run this ad, it's a winner, whatever. [56:31] Will Sartorius: We want to multiply it into different formats. [56:34] Will Sartorius: [VISUAL: Slide titled "Multiply Ads Into Different Formats With AI". It shows the previously generated headline ad on the left, a "Starter Prompt to Convert Your Ad Into Your Other Templates" in the middle, and several different ad formats on the right.] So we have our headline ad. Now, let's assume that we've built other recipe cards. We have more recipes in our cookbook, right? What we can do is we can take our ad that we just created. This is really, really, really cool. This is like probably my biggest breakthrough moment in generating AI ads that I was like, okay, we have something here. You say, I have this winning ad. Here are my other format templates. Maybe it's a statistics ad, maybe it's a a, you know, bullet points, whatever it is. For each format, rewrite the brief to match that template. Same persona, angle, emotion. It's going to rewrite the copy to better fit all of these different versions or variants of the same ad. It's going to run everything through the agents, and then give me the prompts I can utilize in Fal to take this one winning ad or the ad I just generated and turn it into an infinite number of additional ads. So if I have a headline ad and I want to turn it into 14 different versions of different ad types, I can absolutely do that. Alternatively, what I could do is if I had a headline ad and I said, you'll see when you open my my uh headline template, you'll see in the headline template that there are different versions of a headline ad, right? There's one way to, you know, this is just like a very classic header, subheader, CTA. But there's a lot of different ways to do it. Like, some headline ads just have a header. Some headline ads have the product on the right and headline on the left, right? There are, I think I have in the one I'm giving you guys, I have like, you know, four to eight different versions of a headline ad. What you can also say is, take this headline ad, look at my headline template that I have in the folder or the project, and convert this into the other styles in my headline template, and it will give you briefs for that. So like, you are creating a system. And like this is fucking crazy, guys. Like, this is like a world that like never existed before. You have like infinite possibilities here. Like, you can take an ad and convert it into whatever you want. You could say, you know, take this ad and put it into this influencer's hand, or, you know, really, the world is your oyster. So, if you generate and follow these steps, you will get to a place where you can take one ad and multiply it across different formats. You will get to a place where the copy you're writing is crisp, it is emotional, it is persona and angle driven, and you will be finally able to sort of reach the level of volume that Andromeda requires. [59:21] Will Sartorius: [VISUAL: Slide titled "Q&A and/or Live Demo" with a QR code and a link box.] And with that, we'll switch over to Q&A. Uh, appreciate you all. Uh, this is the same QR code I had at the beginning. Um, I was going to offer to do a live demo, but Claude's been down today, so I don't want to necessarily risk risk that. Um, and yeah, so, you know, follow my newsletter. Uh, it comes out every Friday. I give you workflows, I give you prompts. Uh, like I would recommend like if you're going to do anything, that's probably the move. Um, but there's a plenty of other info on uh, on doc. So, yeah, I will uh, hand the keys back over to Evan. [59:53] Evan Lee: Will, you came and smashed the thing. I I want to pick up a bunch of questions, everybody. So hop in there. Will, I wanted to kick off by asking a question that I've seen pop up a little bit here. So there's been some folks who have mentioned like, hey, this seems like things that I can do in Canva, like especially with like just getting started and replacing the headline. And it's like, uh, works similar to that. I'm curious what your response would be to folks like that. [1:00:24] Will Sartorius: Uh, I'm not a designer. I'm innately lazy. And so, I have tried to come up with a system that requires the least amount of work for the most amount of output. Yes, of course, you can do that in Canva. And that's like what, you know, what we're building with Skipper is like you can edit sort of Canva style, uh, directly in platform. Um, but yeah, absolutely. Of course, any of this can be done with Canva. You know, what I might say is like the brief generation process, uh, we still have a human in the loop, like overseeing everything. You know, they're they're going through all these steps. Often times, we'll get a brief and then we'll send that to an editor rather than using, you know, Nano Banana 2. So, this is just a way if you want to sort of volume produce, like that is absolutely a solution. But if you want to just like, you know, generate statics in Canva, at the very least, you now have a system to generate a brief for you and then take that into Canva and execute. [1:01:22] Evan Lee: Well, that's the perfect answer because everybody here, just as a reminder, like on the creative strategy side, a big part of the job is volume. Like you want the volume not just for volume's sake, but you're making sure you're baking in that research, you're breaking in the brand tone, and you're taking insights from winners. So, if I if I can add my two cents there, I think anyone who wants to jump into Canva, like keep doing it. Like if you have the skill set, keep doing it and define what great looks like. But the scaled version of what Will showed, come on. Like that's what anybody would love. So if a designer can just see so many different variations that's been made or new ads that have been made, they can just evaluate them and say like, are they good or not and improve from there. So you built the system in my opinion. [1:02:05] Will Sartorius: Yeah, exactly. [1:02:06] Evan Lee: Okay. Amazing. I'm going to pick up a question now from our from our Q&A. And the first one I'm going to pull up is from Bria and it's a granular one. [1:02:14] Evan Lee: [VISUAL: A question from Bria Mirante is displayed on screen: "Do you give the agents scorecards or do they create them on their own?"] So Bria asks, do you give the agent scorecards or do they create them on their own? [1:02:20] Will Sartorius: Great question. Uh, you can really do it either way. I have the scorecards embedded into the agents, but every time I'm using the copywriting agents, just to be sure, I say, please grade my copy, um, on a scale from 0 to 100 and keep iterating until each agent gives me feedback that is 90 plus, 90 or above out of 100. [1:02:44] Evan Lee: Huge. Love it. Love it. Uh, on the other side of things, I'm I'm wondering basically, I saw a question from Claire pop up. Let me see if I can find it here. [1:02:55] Evan Lee: [VISUAL: A question from Claire Peters is displayed on screen: "Is there a way to force AI tools to use my real product photography and AI generate the background/text/etc? AI is pretty bad at generating product imagery in my category and I have nice product photography already"] Is there a way to force AI tools to use my real product photography and AI generate the background text, etc? And then adding commentary on top. AI is pretty bad at generating product imagery in my category and I have a nice product photography already. [1:03:09] Will Sartorius: Awesome. This is a great question. Yeah, absolutely. So what I would do in that case is I would upload your image into Claude and I would say, you know, I want to generate uh, an ad or whatever it is with the following copy. And then you would upload your brand spec card that has your proper fonts, colors, etc. You would uh, you would update your visual spec card as well. And then you with the prompt that you're generating, you would effectively have it ensure that no aspect of your image changes. Again, like Nano Banana 2 is a really good student. If you tell it what to do, it will do it. So if you say, don't touch any aspect of my product, but please add this header and subheader and this CTA, and I am attaching my visual spec card and my visual style card, it will do that. So absolutely, you can do that. [1:04:08] Evan Lee: Huge. And while we're on the topic of guard rails, another one's popped up from Derek that's climbing up the ranks here. And this is speaking about working with large orgs. [1:04:18] Evan Lee: [VISUAL: A question from Derek Daniels is displayed on screen: "If you're working for a large brand with a defined 'Brand Bible' that includes visual identify, target consumer, design codes, etc., do you recommend uploading the entire 'bible' (which could be well over 100 pages) or breaking it up?"] So if you're working for a large brand with a defined brand bible that includes visual identity, target consumer, design codes, etc., do you recommend uploading the entire bible, which could be well over 100 pages, or breaking it up? [1:04:29] Will Sartorius: Great, great question. And this gets down to like sort of a fundamental principle. Uh, Claude doesn't like PDFs very much. You know, when Claude is looking at a PDF, what it's going to do is it's going to screenshot each page and it's going to, and this is assuming your brand bible is in a PDF or a PowerPoint. Um, what it's going to do is it's going to go and try to screenshot each one. What I would do as a first step is I would upload the PDF into Claude and I would say, can you please convert this to a markdown file? Markdown files are your baseline for everything. .md. Every agent is a markdown file. All the recipe cards, all the templates, they're all markdown files. So, yes, give it the entire thing, but first upload it as a markdown file, and then when you're doing, you know, you're doing the brand execution steps, say, you know, I have this markdown file in the folder, and then you'll be able to have all that properly in there. Um, but yeah, you can you can absolutely do it. Especially if you're working in Claude code, like let's say you built a folder, upload both, you know, upload the the PowerPoint and the PDF and then also upload the markdown file. Like it's going to be scanning through the folder that it's working in and it's going to try, like based on the prompt I gave you, it's going to be trying to build the most comprehensive, uh, you know, spec cards that's humanly possible. [1:05:50] Evan Lee: And then just to just to hit it home for everybody, when you're talking about markdown files and like after interpreting the PDF, are you then separating it out into each section is its own markdown file or does that point doesn't really matter, have it all in one? [1:06:05] Will Sartorius: For the brand guidelines? [1:06:07] Evan Lee: Yes. [1:06:08] Will Sartorius: Um, it doesn't really matter because at the end of the day, what you care the most about is one, like having the information all in one place. Uh, two, having very succinct, uh, spec cards, but it doesn't really matter. And like, unless you have different brand tones for like different products, like I can think of use cases where you'd want a different markdown file. Like maybe you have two arms of the business, you know, one is men's, one is women's, and you talk to them very differently, in which case I would probably have two different markdown files. But if you have a similar tone that you take with all of your customers, then in which case I would say, yeah, absolutely, one is fine. [1:06:47] Evan Lee: Huge. It's a lot simpler than we realize, folks. It's a lot simpler than we realize. [1:06:50] Will Sartorius: And that's the crazy thing. People love to complicate this thing, you know, these things. [1:06:54] Evan Lee: 100%. Well, I'm going to take it a little bit more broad with this next one that's been upvoted a bunch from Ashley. [1:07:02] Evan Lee: [VISUAL: A question from Ashley Humphreys is displayed on screen: "Every time I see AI generated ads, the comments are pretty negative. Do you feel like the efficiencies gained with AI outweigh the negative perception or is this something you use sparingly and rely heavily on creative teams to produce the majority of ads?"] So to read it real quick, every time I see AI generated ads, the comments are pretty negative. Do you feel like the efficiencies gained with AI outweigh the negative perception or is this something you use sparingly and rely heavily on creative teams to produce the majority of ads? [1:07:16] Will Sartorius: Okay, first off, great question again. Um, first thing I would say is like, if I showed you guys any of the statics I created, would you know they were AI generated? The answer is probably not. So, as it [1:07:26] Evan Lee: Let us know in the chat. Yeah, let us know in the chat what you all think. [1:07:29] Will Sartorius: As it pertains to statics, it's really hard. We've gotten to the point, like again, if you follow my system, you'll be able to generate ads that are on brand. Um, when it comes to videos, I think it's probably we've we've tested a lot. You know, you see on Twitter like these Play-Doh ads or like the, you know, Pixar things, like, see, it feels a little drop shippy and sort of gross to me. We also don't use AI UGC because I just like I'm like sort of like morally opposed to it. Um, when it comes to sort of AI video, I think if you really lean into it and make it like, this is an AI video and like don't try to hide it, then I think that is a much safer bet. You will get a lot of comments. I so the the first time I did one of the Pixar ads for one of our clients, they were they're an NA beer company. They had like a an Irish beer and I did like a little, uh, leprechaun and it was like, you know, they keep stealing my beer, you know, uh, like the Lucky Charms guy. And, uh, we got a lot of, we got a lot of, you know, comments saying, hey, you know, I'm never going to buy from this company. It's AI slop. What I ended up doing is I ended up like sort of looking at those people and like cross-referencing them on Shopify, and these people were never going to purchase in the first place. They are just there to sort of troll, and that ad, the client first reached out to us and be like, hey, this is a lot of negative comments, like, what do you guys think? And we're like, I don't know, it's printing, like it's working really, really well. So, yeah, that's sort of one, like you sort of have to have a bit of a stomach for it. You can also delete the comments or hide them. The second solution, which is a much better one, is spin up a Facebook page, uh, you know, and white list it not from your own account. So maybe it's like a Facebook page that's called, uh, you know, Irish Beer Club, and then you run the ad from that account. It's still in your business manager, but at the end of the day, like the brand itself isn't taking the heat, um, if you don't have the stomach for it, but these ads do work, especially for, I hate to say it, but like for older demos, like the sort of the AI video ads work pretty well. [1:07:57] Evan Lee: That was a gem in there, everybody, just on the second page. Like that was, you said it very casually, but a big gem in there for everybody. [1:08:07] Will Sartorius: Slop is slop, whether it's AI or human-made at that point, too. [1:08:11] Evan Lee: It's so true. It's so true. It's so true. Well, I got I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about like the tool stack. So let's pull a Derek question up here. [1:08:22] Evan Lee: [VISUAL: A question from Derek Daniels is displayed on screen: "Assuming one has the most rudimentary access to tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad?"] Assuming one has the most rudimentary access to tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:08:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:09:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:09:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:09:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:09:46] Will Sartorius: I don't hear anything. [1:09:49] Evan Lee: You got me, man. You got me. Oh, man. [1:09:52] Will Sartorius: I'm just kidding. [1:09:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:00:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:03:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. So it's like so true. It's so true. It's so true. Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:03:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:04:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:04:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:04:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:04:46] Will Sartorius: I don't hear anything. [1:04:49] Evan Lee: You got me, man. You got me. Oh, man. [1:04:52] Will Sartorius: I'm just kidding. [1:04:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:05:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:08:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:08:11] Will Sartorius: It's so true. It's so true. It's so true. [1:08:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:08:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:09:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:09:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:09:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:09:46] Will Sartorius: I don't hear anything. [1:09:49] Evan Lee: You got me, man. You got me. Oh, man. [1:09:52] Will Sartorius: I'm just kidding. [1:09:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:10:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:13:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:13:11] Will Sartorius: It's so true. It's so true. It's so true. [1:13:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:13:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:14:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:14:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:14:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:14:46] Will Sartorius: I don't hear anything. [1:14:49] Evan Lee: You got me, man. You got me. Oh, man. [1:14:52] Will Sartorius: I'm just kidding. [1:14:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:15:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:18:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:18:11] Will Sartorius: It's so true. It's so true. It's so true. [1:18:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:18:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:19:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:19:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:19:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:19:46] Will Sartorius: I don't hear anything. [1:19:49] Evan Lee: You got me, man. You got me. Oh, man. [1:19:52] Will Sartorius: I'm just kidding. [1:19:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:20:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:23:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:23:11] Will Sartorius: It's so true. It's so true. It's so true. [1:23:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:23:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:24:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:24:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:24:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:24:46] Will Sartorius: I don't hear anything. [1:24:49] Evan Lee: You got me, man. You got me. Oh, man. [1:24:52] Will Sartorius: I'm just kidding. [1:24:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:25:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:28:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:28:11] Will Sartorius: It's so true. It's so true. It's so true. [1:28:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:28:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:29:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:29:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:29:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:29:46] Will Sartorius: I don't hear anything. [1:29:49] Evan Lee: You got me, man. You got me. Oh, man. [1:29:52] Will Sartorius: I'm just kidding. [1:29:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:30:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:33:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:33:11] Will Sartorius: It's so true. It's so true. It's so true. [1:33:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:33:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:34:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:34:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:34:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:34:46] Will Sartorius: I don't hear anything. [1:34:49] Evan Lee: You got me, man. You got me. Oh, man. [1:34:52] Will Sartorius: I'm just kidding. [1:34:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:35:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:38:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:38:11] Will Sartorius: It's so true. It's so true. It's so true. [1:38:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:38:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:39:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:39:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:39:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:39:46] Will Sartorius: I don't hear anything. [1:39:49] Evan Lee: You got me, man. You got me. Oh, man. [1:39:52] Will Sartorius: I'm just kidding. [1:39:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:40:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:43:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:43:11] Will Sartorius: It's so true. It's so true. It's so true. [1:43:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:43:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:44:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:44:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:44:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:44:46] Will Sartorius: I don't hear anything. [1:44:49] Evan Lee: You got me, man. You got me. Oh, man. [1:44:52] Will Sartorius: I'm just kidding. [1:44:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:45:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:48:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:48:11] Will Sartorius: It's so true. It's so true. It's so true. [1:48:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:48:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:49:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:49:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:49:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:49:46] Will Sartorius: I don't hear anything. [1:49:49] Evan Lee: You got me, man. You got me. Oh, man. [1:49:52] Will Sartorius: I'm just kidding. [1:49:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:50:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:53:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:53:11] Will Sartorius: It's so true. It's so true. It's so true. [1:53:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:53:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:54:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:54:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:54:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:54:46] Will Sartorius: I don't hear anything. [1:54:49] Evan Lee: You got me, man. You got me. Oh, man. [1:54:52] Will Sartorius: I'm just kidding. [1:54:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [1:55:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [1:58:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [1:58:11] Will Sartorius: It's so true. It's so true. It's so true. [1:58:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [1:58:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [1:59:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [1:59:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [1:59:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [1:59:46] Will Sartorius: I don't hear anything. [1:59:49] Evan Lee: You got me, man. You got me. Oh, man. [1:59:52] Will Sartorius: I'm just kidding. [1:59:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:00:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:03:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:03:11] Will Sartorius: It's so true. It's so true. It's so true. [2:03:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:03:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:04:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:04:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:04:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:04:46] Will Sartorius: I don't hear anything. [2:04:49] Evan Lee: You got me, man. You got me. Oh, man. [2:04:52] Will Sartorius: I'm just kidding. [2:04:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:05:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:08:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:08:11] Will Sartorius: It's so true. It's so true. It's so true. [2:08:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:08:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:09:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:09:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:09:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:09:46] Will Sartorius: I don't hear anything. [2:09:49] Evan Lee: You got me, man. You got me. Oh, man. [2:09:52] Will Sartorius: I'm just kidding. [2:09:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:10:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:13:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:13:11] Will Sartorius: It's so true. It's so true. It's so true. [2:13:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:13:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:14:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:14:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:14:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:14:46] Will Sartorius: I don't hear anything. [2:14:49] Evan Lee: You got me, man. You got me. Oh, man. [2:14:52] Will Sartorius: I'm just kidding. [2:14:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:15:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:18:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:18:11] Will Sartorius: It's so true. It's so true. It's so true. [2:18:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:18:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:19:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:19:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:19:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:19:46] Will Sartorius: I don't hear anything. [2:19:49] Evan Lee: You got me, man. You got me. Oh, man. [2:19:52] Will Sartorius: I'm just kidding. [2:19:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:20:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:23:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:23:11] Will Sartorius: It's so true. It's so true. It's so true. [2:23:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:23:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:24:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:24:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:24:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:24:46] Will Sartorius: I don't hear anything. [2:24:49] Evan Lee: You got me, man. You got me. Oh, man. [2:24:52] Will Sartorius: I'm just kidding. [2:24:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:25:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:28:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:28:11] Will Sartorius: It's so true. It's so true. It's so true. [2:28:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:28:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:29:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:29:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:29:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:29:46] Will Sartorius: I don't hear anything. [2:29:49] Evan Lee: You got me, man. You got me. Oh, man. [2:29:52] Will Sartorius: I'm just kidding. [2:29:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:30:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:33:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:33:11] Will Sartorius: It's so true. It's so true. It's so true. [2:33:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:33:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:34:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:34:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:34:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:34:46] Will Sartorius: I don't hear anything. [2:34:49] Evan Lee: You got me, man. You got me. Oh, man. [2:34:52] Will Sartorius: I'm just kidding. [2:34:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:35:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:38:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:38:11] Will Sartorius: It's so true. It's so true. It's so true. [2:38:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:38:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:39:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:39:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:39:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:39:46] Will Sartorius: I don't hear anything. [2:39:49] Evan Lee: You got me, man. You got me. Oh, man. [2:39:52] Will Sartorius: I'm just kidding. [2:39:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:40:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:43:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:43:11] Will Sartorius: It's so true. It's so true. It's so true. [2:43:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:43:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:44:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:44:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:44:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:44:46] Will Sartorius: I don't hear anything. [2:44:49] Evan Lee: You got me, man. You got me. Oh, man. [2:44:52] Will Sartorius: I'm just kidding. [2:44:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:45:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:48:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:48:11] Will Sartorius: It's so true. It's so true. It's so true. [2:48:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:48:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:49:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:49:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:49:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:49:46] Will Sartorius: I don't hear anything. [2:49:49] Evan Lee: You got me, man. You got me. Oh, man. [2:49:52] Will Sartorius: I'm just kidding. [2:49:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:50:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:53:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:53:11] Will Sartorius: It's so true. It's so true. It's so true. [2:53:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:53:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:54:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:54:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:54:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:54:46] Will Sartorius: I don't hear anything. [2:54:49] Evan Lee: You got me, man. You got me. Oh, man. [2:54:52] Will Sartorius: I'm just kidding. [2:54:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [2:55:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [2:58:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [2:58:11] Will Sartorius: It's so true. It's so true. It's so true. [2:58:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [2:58:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [2:59:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [2:59:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [2:59:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [2:59:46] Will Sartorius: I don't hear anything. [2:59:49] Evan Lee: You got me, man. You got me. Oh, man. [2:59:52] Will Sartorius: I'm just kidding. [2:59:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:00:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:03:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:03:11] Will Sartorius: It's so true. It's so true. It's so true. [3:03:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:03:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:04:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:04:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:04:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:04:46] Will Sartorius: I don't hear anything. [3:04:49] Evan Lee: You got me, man. You got me. Oh, man. [3:04:52] Will Sartorius: I'm just kidding. [3:04:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:05:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:08:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:08:11] Will Sartorius: It's so true. It's so true. It's so true. [3:08:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:08:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:09:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:09:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:09:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:09:46] Will Sartorius: I don't hear anything. [3:09:49] Evan Lee: You got me, man. You got me. Oh, man. [3:09:52] Will Sartorius: I'm just kidding. [3:09:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:10:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:13:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:13:11] Will Sartorius: It's so true. It's so true. It's so true. [3:13:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:13:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:14:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:14:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:14:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:14:46] Will Sartorius: I don't hear anything. [3:14:49] Evan Lee: You got me, man. You got me. Oh, man. [3:14:52] Will Sartorius: I'm just kidding. [3:14:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:15:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:18:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:18:11] Will Sartorius: It's so true. It's so true. It's so true. [3:18:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:18:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:19:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:19:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:19:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:19:46] Will Sartorius: I don't hear anything. [3:19:49] Evan Lee: You got me, man. You got me. Oh, man. [3:19:52] Will Sartorius: I'm just kidding. [3:19:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:20:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:23:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:23:11] Will Sartorius: It's so true. It's so true. It's so true. [3:23:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:23:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:24:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:24:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:24:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:24:46] Will Sartorius: I don't hear anything. [3:24:49] Evan Lee: You got me, man. You got me. Oh, man. [3:24:52] Will Sartorius: I'm just kidding. [3:24:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:25:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:28:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:28:11] Will Sartorius: It's so true. It's so true. It's so true. [3:28:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:28:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:29:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:29:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:29:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:29:46] Will Sartorius: I don't hear anything. [3:29:49] Evan Lee: You got me, man. You got me. Oh, man. [3:29:52] Will Sartorius: I'm just kidding. [3:29:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:30:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:33:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:33:11] Will Sartorius: It's so true. It's so true. It's so true. [3:33:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:33:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:34:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:34:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:34:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:34:46] Will Sartorius: I don't hear anything. [3:34:49] Evan Lee: You got me, man. You got me. Oh, man. [3:34:52] Will Sartorius: I'm just kidding. [3:34:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:35:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:38:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:38:11] Will Sartorius: It's so true. It's so true. It's so true. [3:38:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:38:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:39:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:39:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:39:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:39:46] Will Sartorius: I don't hear anything. [3:39:49] Evan Lee: You got me, man. You got me. Oh, man. [3:39:52] Will Sartorius: I'm just kidding. [3:39:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:40:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:43:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:43:11] Will Sartorius: It's so true. It's so true. It's so true. [3:43:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:43:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:44:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:44:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:44:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:44:46] Will Sartorius: I don't hear anything. [3:44:49] Evan Lee: You got me, man. You got me. Oh, man. [3:44:52] Will Sartorius: I'm just kidding. [3:44:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:45:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:48:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:48:11] Will Sartorius: It's so true. It's so true. It's so true. [3:48:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:48:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:49:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:49:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:49:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:49:46] Will Sartorius: I don't hear anything. [3:49:49] Evan Lee: You got me, man. You got me. Oh, man. [3:49:52] Will Sartorius: I'm just kidding. [3:49:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:50:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:53:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:53:11] Will Sartorius: It's so true. It's so true. It's so true. [3:53:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:53:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:54:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:54:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:54:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:54:46] Will Sartorius: I don't hear anything. [3:54:49] Evan Lee: You got me, man. You got me. Oh, man. [3:54:52] Will Sartorius: I'm just kidding. [3:54:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [3:55:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [3:58:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [3:58:11] Will Sartorius: It's so true. It's so true. It's so true. [3:58:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [3:58:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [3:59:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [3:59:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [3:59:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [3:59:46] Will Sartorius: I don't hear anything. [3:59:49] Evan Lee: You got me, man. You got me. Oh, man. [3:59:52] Will Sartorius: I'm just kidding. [3:59:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [4:00:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [4:03:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [4:03:11] Will Sartorius: It's so true. It's so true. It's so true. [4:03:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use to get to an MVP for an ad? So it's like a lean tool stack to to implement what you've talked about. [4:03:34] Will Sartorius: Sweet. Uh, Claude chat, uh, and Gemini. That's all you need. Yeah, you can you can even have the free version of Claude chat. You don't need to use the, I mean, Opus 4.6 is going to give you the best results, but you don't need to use the best Claude model. You can use the free Claude version and then you just you work directly in Gemini and Gemini, you don't have to pay for. So, I mean, you can do this on pennies if if nothing at all. So, yeah, it's a very cheap stack. Um, and, uh, you know, anyone can really do it. [4:04:03] Evan Lee: Will, you killed it from the basics to how to scale this thing. It was absolutely incredible, man. Do you have any final words to leave with our audience? [4:04:10] Will Sartorius: Um, yeah, thanks for hopping on. Um, again, you know, I give away a lot of this stuff on my Friday newsletter. Um, and, you know, if you want beta access to our, uh, you know, image generation platform, you know, feel free to sign up for that as well. But appreciate you all. Uh, thanks for signing on and, uh, you know, chatting with me from Brooklyn. [4:04:32] Evan Lee: Will, you got to go check the chat after this. They're going nuts. So you got a lot to read through. There's thousands of messages, but you crushed it. Everybody, throw love in the chat. Throw your favorite emoji in there to talk about how you're feeling. But give a round of applause for Will, everybody. Give a round of applause for Will. [4:04:46] Will Sartorius: I don't hear anything. [4:04:49] Evan Lee: You got me, man. You got me. Oh, man. [4:04:52] Will Sartorius: I'm just kidding. [4:04:52] Evan Lee: The interesting, the interesting thing on the negative comment side too that I'm I'm curious to get your perspective on is like after you cross-referenced and saw who was commenting versus what you saw on Shopify, how much of that pulled back to your, like, research phases? So it's like understanding your buyer and then saying, I know they'll probably like this, ultimately. Like what was that connection? [4:05:13] Will Sartorius: Yeah, of course. So, uh, it's a great question. Like, this brand is like very, very tongue in cheek. Uh, you know, uses curse words with like, you know, uh, crossed out, uh, you know, the U and FUCK, right? Um, and so we know the persona pretty well at this point. Like we know that if we're like sort of cheeky and edgy and we're not trying to hide anything, you know, it's a very like much like a Liquid Death vibe. Um, and so we have a pretty good sense of the consumer. Um, but yeah, I mean, listen, at the end of the day, this entire presentation assumes that you know baseline creative strategy. It assumes you know which persona to target. It assumes you know which angle to focus on. It assumes you know which emotion to focus on. You know, I I didn't give you that. I just gave you the system after you've done all of your creative strategy and actually how to execute. So like, there is this mushy human element that is really, really integral to this entire process. And again, like, you know, people like AI creative guy, like, listen, 90% of the ads, even more, are generated from like human thought, right? We're not just like churning out shit for the sake of churning out shit, right? So I shouldn't swear. Uh, my wife would get mad at me. Uh, the, uh, the idea is like there is this human element. Like I need to know which persona, angle, emotion I need to target, right? And which product will work best with that. And like, the AI is not deciding that for us. You know, the the the creative strategists are. And one last piece to this because I don't want to get like too too deep into a rant. You need to hit volume. That's where we're in this day and age where meta requires volume. So, we have used AI to help supplement that volume part of the equation. You also really need to have human elements, human ads. Like, I was just in Lisbon the other week giving a presentation for Foxwell, and, you know, I was chatting with Harry Dalmage, probably butchering his last name, and we we chatted for a while about, like, I'm doing AI and he's doing these like really good human two to four minute ads that are just like, there's no way AI could replicate that. So like, you need to have both. No ad account is going to take off with just statics and gifts, unless you're that, you know, Medby Medby guy who's just like, you know, spitting up fake doctors, which is problematic in and of itself. Don't do that. Don't do that. Um, but you need to have very human-oriented content as well, especially for top of funnel in order to educate. [4:08:05] Evan Lee: Such a good reminder, Will. And honestly, slop is slop, whether it's AI or human-made. At that point, too. [4:08:11] Will Sartorius: It's so true. It's so true. It's so true. [4:08:13] Evan Lee: Well, I got one more question before I get you out of here. So let's bring it up. Everyone's been asking about the tool stack. So let's pull a Derek question up here. Assuming one has the most rudimentary access to the tools, what basic, accessible, and cheap tools and process would you use