AI Secrets for Ad Research 2025: Old Way vs New Way (Complete Guide)

Alex Cooper: We're going to be covering today how to use AI to conduct winning creative research in 2020 2025.

Alex Cooper: 2025, not 2020 2025. Uh, and we're going to make a case for why creative strategists should be upskilling into the creative strategist prompt engineer role. Now, a quick intro for those of you who don't know who we are.

Alex Cooper: Uh, we are the co-founders of a company called Human Squared. We work with enterprise brands such as Unilever and Virgin, uh, to build them AI workflows and teach them the stuff that we're going to be teaching you today. Uh, we also make content helping you make better direct response ads with the help of AI. Some of you might also recognize me as the co-founder of Ad-Crate, uh, a performance creative agency we work with eight and nine figure DC brands to make them direct response ads using, uh, the help of AI.

So, guys, we have a lot that we want to get through today. Uh, so we better crack on.

Alex Cooper: Uh, let's, uh, have a look at the agenda. So, what Jimmy and I want to talk through is why most audience research will be done by AI and what you can do about it. Uh, the case for the creative strategist prompt engineer and why we think it will thrive this year. The exact tools we use at Ad-Crate, um, to conduct research across 30 brands. Key mistakes advertisers make with AI. Uh, we're going to share some automations and some of the workflows that we've built internally, uh, which is going to be super fun. And finally, how to kind of put this all together and incentivize AI usage across your team. And if you stick around till the end, we have a very special giveaway. It's the biggest giveaway we've ever done. Um, a lot of workflows, a lot of documents, a lot of prompts. So make sure you do not miss that.

Alex Cooper: I want to start by, uh, talking about why I think this is so important. And Evan kind of alluded to it in, uh, the, um, uh, in the intro. Uh, I personally do not believe that creative strategists will be replaced. If you are to be replaced, you'll be replaced by another creative strategist using AI better than you. Uh, the way the industry is going, I just think there's, um, like there's a lot of information out there on like on Twitter or LinkedIn saying like creative strategists are going to be replaced. I just think the role is going to change and we're going to cover, uh, what we think is going to change and how you can adapt into that today.

Jimmy Slagle: Yeah, yeah, yeah. Um, for sure. And honestly, uh, I think, I think where this industry is moving is really creative strategists are are going to become, uh, this this term that we're going to use a lot, which is creative strategist prompt engineer, which means that we think most, uh, of the work that creative strategists do today and even outside of creative strategy, whether it be design, video editing, video generation is all just going to be done using prompts. And so, I can show you what what we mean by that.

Jimmy Slagle: Um, so the first part of this research, this is obviously the meat of what we're going to be talking about. This is just a small sample size. Um, but right here, this is the ability to already conduct research without needing to know how to code. That's the greatest part is you can go into a tool, um, like Gumloop or or Motion is going to get there as well, to where you can talk with Reddit as if it was natural language. So, hey, what are people saying about my product or what are people saying about my pain points, uh, on Reddit or, uh, you know, YouTube or Twitter or Facebook ad libraries. All of these things you're just going to be able to have natural conversations with instead of you having to go and look at Reddit and scroll through for hours to try to find, uh, gold nuggets or go to a tool like Foreplay or a tool, uh, like the actual ad library itself to look at different ads. Everything's going to be a prompt. You can just ask what's happening here, what's happening there.

Jimmy Slagle: Um, but beyond that, uh, we think like, if you guys are familiar with ChatGPT image gen, these two images right here were created using a prompt. We didn't touch Photoshop, we didn't touch Canva, we didn't touch anything. This was just from going into ChatGPT image gen, having a pretty advanced prompt that was able to get us something of this quality, which is which is absolutely crazy.

Jimmy Slagle: Um, and then beyond there too, landing pages. This same thing. I I I went to ChatGPT image gen, had it come up with some pretty impressive landing pages, um, uh, that you can then go and use. And so, these these are two examples of where just natural language was able to create something this good. And we're so early in in the world of AI. The fact that ChatGPT didn't even exist, you know, whatever it was, three years ago is is is just wild to think about where the next three years are going to go.

Jimmy Slagle: Um, even video, if you guys are on social media, you may have seen, um, uh, a tool called, uh, Veo, which is from Google. We're going to play this, turn off the sound so you can just watch, but everything in this video is completely AI generated. Um, and so this is someone just explaining what they wanted to see within a video, and this is a minute and a half montage of all completely AI generated clips, um, which is pretty crazy from a creative, uh, strategist perspective because a lot of our role is is figuring out what should we create and then having to coordinate all the shoots, all the creators, um, all the, uh, editing, everything there. But the future of this world is really just going to be you describing to a tool like Veo what you want generated, and then something, um, of this quality is able to be created, which, which is pretty, pretty impressive.

Jimmy Slagle: Um, and maybe it's not video, uh, generation itself, but for sure editing is going to go to a prompt first world. So this is another tool that I can show you where this is essentially an example of someone uploading a bunch of clips and then just describing the editing styles that they want. Um, and again, this is a huge unlock for creative strategists. We all know what we want in our mind, we just don't want to spend the hours or learn how to use Premiere Pro or CapCut to get there. A tool like this where it's just natural language video editing, um, is going to come out and and these tools are going to get better and better over time as well.

Jimmy Slagle: Um, but I know, I know the the core of this is going to be, um, uh, research. So, let's just start there. Uh, that stuff will come, come later.

Alex Cooper: Yeah, and we have an internal saying at Ad-Crate that 90% of your creative problems can be traced back to not doing a good enough research. I actually think it's more than that. Uh, it's the most impactful part of the ad creation process. Um, to put it simply, bad research equals bad ads.

Alex Cooper: There we go. Um, and we say good research like a lot of the time. Like I I I think that a lot of people think that, you know, good research is just checking boxes and saying like, you know, I've done, I've looked at my customer reviews, I've looked at my comments, I've looked at my, I've looked at Reddit, for example.

Alex Cooper: Um, but there's a difference between like conducting research and conducting good research. Conducting good ad like creative strategy research is about finding those little golden nuggets, finding the words and phrases that your customers are actually using when they're talking about your pain point or your desired transformation. For example, Obvi, when they, uh, when they found people talking about FUPA, uh, I noticed it was way less noticeable and my jeans started fitting better. Uh, and then take taking that inside phrase that only someone who struggles with that problem, uh, would understand, putting on a static, and then, I mean, Ash tweeted that that was one of their best performing ads. It's just about creating ads that make people think like, there's no way they can't be one of us. Like, how do they know about this phrase or this, you know, this wording or like this concept?

Alex Cooper: Um, and that only comes from real creative strategy research. So this is what it means to be a creative strategist. Uh, it's finding new use cases. So it's it's going onto Reddit posts, uh, that was made seven years ago where someone was talking about, you know, I actually use my anti-dandruff shampoo for my beard as well. Uh, and it's taking that and then putting that on, uh, our, um, our statics and saying that this is not just for your scalp, for your beard. And that worked inside the ad account as well. It's really like deeply listening to our customers and the way that they communicate about their pain point or about their desired transformation.

Alex Cooper: Now, the problem is, it's really difficult to do that across all of these platforms at that level. Do that across the ad account, across trends, Reddit, uh, reviews, Quora, Pinterest, Instagram, like you can put them all up there. It's really difficult to get that level of good creative strategy research by doing it across all of these different mediums.

Alex Cooper: At least if you're only doing it on your own.

Alex Cooper: And like this is where Jimmy and I want to introduce the concept of a creative strategist prompt engineer. And we really think that this is the way that the industry is going. Right now, a creative strategist is trying to do all of these things to inform which ads that they're going to produce, whereas the creative strategist of the future, we believe, are, you know, recognizing that this is a process that, you know, even though research is very time consuming, it is also a process that is repeatable.

Alex Cooper: And if the process is repeatable,

Alex Cooper: uh, then that means it can be systemized.

Alex Cooper: And if it can be systemized, then that means it can be done with AI.

Alex Cooper: And that's the whole concept we want to, uh, get across today. Um, and the reason for why we think, uh, people should be transitioning into, uh, creative strategist prompt engineers. Brands should be hiring for this role, not in those words, but instead of, you should only be hiring creative strategists who are AI first. Agencies should be upskilling their teams into this role, or their strategists into this role. And strategists, you guys need to start evolving into this role as soon as possible. And the cool thing is because we are so relatively early of people who are actually using this throughout their entire workflow properly, there is a big opportunity, uh, for people to get ahead now, whether you're a strategist, a freelancer, an agency, or a brand.

Jimmy Slagle: And, and so just giving you guys a little glimpse of like what we think this is going to look like from a very, uh, tangible and and and digestible, uh, format.

Jimmy Slagle: Um, honestly, we think that eventually everyone here is going to have their own AI research agent. Um, and essentially what this is going to be able to do is you're going to have an agent that can go and scrape Reddit, you're going to have an agent that can go and scrape organic TikTok, you're going to have an agent that's going to scrape your your ad account, um, all of these different things below you, and they're just going to do all of that for you and give you that analysis.

Jimmy Slagle: And and and primarily, we think a lot of this could happen within Slack. So, for example, you could set up an agent that's essentially, uh, goes and and scrapes, uh, your customer reviews, it could scrape Reddit, it could look at your post-purchase surveys, support tickets, all of these different, uh, data sources and say, hey, you know, good morning, Alex, we looked through all of these different things, and these are like the unique insights that we found that you guys haven't really, uh, uh, focused on from a, from an ad perspective yet, that could be interesting, or just more, more personas to think about. You're just going to have these agents working 24/7, uh, trying to find these unique insights, which is something that just humans won't necessarily be able to do at scale, as Alex talked about.

Jimmy Slagle: And and so that's why we don't think that this should scare you. This is not going to replace the creative strategist. We think it's just going to allow creative strategists to, uh, make ads faster, and it's going to be a lot easier. You're not necessarily going to need the the, uh, level of design work that you did before or the amount of video editors that you need before or the amount of creators that you needed. But the creative strategist, it's kind of your, the the world is your oyster, especially when it comes to research. Um, a lot of your time was probably spent on research, or at least it should be, and now that is just going to be be the opposite. You get to spend a lot of your time thinking of all those ideas. And then the word everyone wants to hear these days, it's going to be way cheaper to make ads, um, as tools like image gen and as tools like Veo, uh, come come around and get better and better.

Jimmy Slagle: Um, the cost of creating ads is almost going to go to zero. And and just again, that concept of knowing what to create is going to be the most valuable thing. So, what what do we think this means just from like the macro perspective? We think the AI is really just going to raise the floor. Um, there's not going to be a thing as like a bad creative strategist or or a decent creative strategist because we think that AI is going to allow everyone to just do so much more. And so the people that are really going to thrive are those that, um, um, take this this this skill and this art of of creative strategy seriously and spend time on webinars like this trying to just learn more and more about this this craft. Um, so again, that's that's where, uh, uh, we think AI is going to still struggle. We don't think it's going to raise the ceiling all that much. Um, we just think that the floor is going to be way higher. So, you're going to hear this a lot from us. The value is going to come from knowing what to create, not necessarily the skill of creating ads itself.

Jimmy Slagle: Um, so, how is a creative strategist prompt engineer going to add value? We think it's the quality of direction, so just that overall strategy and focus, um, still, you know, knowing what platforms to focus on and the strategies within each of those. The ideas, I think this is going to be the biggest one. Everyone here has a different taste, uh, which is, you know, their kind of spin-off of what a good ad is or or, um, ad styles that they really enjoy and and think are valuable. That's going to be what's unique about you. And then again, this this new concept of starting to work with these agents. Do you know how to communicate with these agents? Do you know how to, uh, set them up? And don't worry too much about that right now. Uh, not not to to overwhelm you, but that is going to be something that over time, people are just going to have to become more and more familiar with. And then finally, just again, that quality of the context. Um, this is something that we're super passionate about. If you ask ChatGPT to create a good ad for you right now, uh, it's not going to know because it's been trained off of thousands of blogs that exist that are only out there for SEO purposes on how to create better Facebook ads in 2024, 2025. It's not actually looking at ads to be able to tell you like, hey, this is what is going to be, uh, worth testing or this is how you create a good ad. And so this is where you need to almost train it on like, hey, here are 10 different static ads that we've run. Um, can you help me think of another idea? You're going to start to see ChatGPT be way more valuable when you start to use it like that and when you give it really good examples versus just generically asking it to create an ad.

Alex Cooper: So, all that said, we think that the creative strategy job is going to become even more valuable because, you know, this this world is going to a prompt first world and they're just going to be able to do so much more and provide so much more value.

Alex Cooper: So, you might be watching this and thinking like, that's great. Uh, I appreciate that like that's the way the industry is going. Um, but the reality is, I still need to make ads today. Um, so what does this mean for me today?

Alex Cooper: Let's go through a few things that you can do to start moving into that creative strategist prompt engineer role and give you practical things, uh, that you can apply, um, with AI, uh, into your workflow today. Um, so, four things we're going to go through. Uh, adding AI tools to your existing tech stack, becoming a good prompter, um, building our workflows, and then finally putting it all together and thinking about how we can scale it, not just across the creative team, but actually across the whole business. We want everyone here to become an AI champion, uh, and add value to their organization.

Alex Cooper: So, let's start with, uh, adding AI tools into your tech stack. And before I, uh, get into this, I do just want to say, I think a lot of people tend to, uh, try and find like the one tool that's going to replace their entire workflow and almost operate as if like it's an infinite ads machine. Um, and the reality is, like, I mean, Jimmy and I have pretty much tried every tool out there. They're not there today. Uh, and we think a lot more of a defensible approach is to look at your existing process and think about what already works and then find areas of that in which AI can, uh, enhance. So, for example, we spoke about Reddit earlier. If you already manually go and crawl Reddit for posts about dandruff in this example, you can use Reddit's, uh, AI chat function, Reddit Answers. And I've started tweeting about this recently, guys. It came out about a month ago. Um, it is so good. Genuinely, like everyone here, uh, I believe will find value from Reddit Answers. I used to use Gigabrain, um, which tapped into the Reddit API and gave you answers. I don't know whether it's because it's trained on local data here, but the outputs from this are just so much stronger. And like we've got so many nuggets that are like otherwise difficult to find in Reddit because sometimes in Reddit, like you don't know what you need to search to get to the juicy stuff.

Alex Cooper: But this stuff, you ask the question and it gets you straight there. And like some of these are like ad-ready headlines or ad-ready concepts. So, you know, I asked what emotive language are people using to describe dandruff? I want real words from real people. It gave me a load of different phrases that real people are saying about the pain point of dandruff from multiple different Reddit threads, and it references those Reddit threads if I wanted to click into them and see the exact words and the comments and the replies what people are saying. And the cool thing is, I took that question and I said to ChatGPT, this question gave me a really good response on Reddit Answers. Give me like 20 other questions that I could ask, uh, Reddit, um, to help me find out about my pain point and then standardize them. So, I believe we're going to be giving these away at the end. Uh, these are just a few of them, uh, that I enjoy, uh, asking Reddit. Uh, again, um, just to highlight a couple of them that I I I really, uh, use often. Uh, this one, number five, inside jokes, nicknames, or phrases. This is the example that I showed from Obvi earlier. Um, that is like, that is how you create ads that make people think there's no way they're not one of us. Like, how do they know about this phrase or this, you know, this wording or like this concept? I like, what do you type into Google when you're looking for help with problem as well? Just just to see what like what search terms people would be using, uh, in their words, uh, to describe their problem. And what's a funny or embarrassing story? Sometimes if you have, you know, specific stories, again, they feel super relatable and authentic, as we've been able to build ads out of that as well.

Jimmy Slagle: The next one is deep research. If you guys haven't used ChatGPT deep research, this one is a really powerful tool, especially when you can get the prompts right for it. Um, what deep research does is it usually takes around eight to 15 minutes to just go through all the different resources that it can find on the internet. So let's just say, uh, you know, you're, you're, um, the perfect jean, and you ask the question of like, hey, I want you to find all of the the buyer personas for the perfect jean based on what you're seeing across different blogs or YouTube videos or whatever it is, um, where people are talking about like what jeans to purchase. And then ChatGPT is going to go and literally spend eight minutes just crawling the internet, uh, uh, reasoning with itself of like, oh, this would be something, or or maybe this is a unique persona that that we could target. And then it's going to give you just this massive report on everything that it found with the links back to, uh, the reference material. So, ChatGPT deep research is is seriously one of my favorite tools for research, um, especially when you can get creative with the prompts.

Jimmy Slagle: And, uh, again, this is part of the giveaway. If you guys stay till the end, we're going to be giving away 20 deep research prompts that you can just copy into ChatGPT deep research, um, enter in your URL, um, and you're going to be able to use the prompts that we're using internally, which, which, um, you're going to get a lot of fun, uh, and and unique answers for. These are questions that, um, are definitely more, uh, creative and not obvious. So I'd be surprised if if, uh, many people have tested them out.

Alex Cooper: Next tool I want to talk about is Poppy. And again, this is another one that I shout a lot about. People think I'm on the Poppy sales team. Uh, I'm not. I don't get paid by them, unfortunately. Um, but it is a very, very good tool. Um, what Poppy is is an AI wrapper, and that means that, you know, you can tap into different models like Claude or GPT or whatever you want to use and, uh, chat with it the same way you would. But the difference with Poppy is that you can add additional context to those chats that you wouldn't otherwise be able to add natively. So you can add things like, uh, YouTube videos, TikTok videos, Instagram videos, uh, PDFs, um, voice messages, text, uh, websites. There's a bunch of different inputs that you can give it. Uh, some of which you can add directly to a GPT, but, uh, some of which you can't.

Alex Cooper: And the reason I really like Poppy is it gives you this like visual, um, mind map style like UI. If you just go back to the previous slide for me, Jimmy, I just want to explain this real quick. Um, this is a board that I built inside of Poppy that is like a, what we call a copycat, uh, workflow, uh, whereby, like, I have given it, it's too small in here, but like, you know, this is part of the giveaway as well, so you'll be get get access to this board. We've given it a document on how to create static ads. Like everything that we know about static ads, here's how to create a good static ad. We've given it context about our brand, about our product, and, um, about like what the offer is. And then I've just got this like little group at the bottom there where I can just drop any static or video ad in and it will create a brief for me to create that static ad. Now, it doesn't do generation yet, but I'm assuming with the image gen API, uh, being released, then it won't be long until it does. But now anytime I see a static ad like on Twitter or in an ad library or wherever, I just drop it into here and it will give me ideas of headlines I can spin up in the style of the static ad or video ad that I, uh, that I dropped into here. So, um, Poppy is a super useful tool. Jimmy actually took that one step further and built this, uh, monster of a Poppy board, uh, where we have 150, uh, static ads that you can choose which one you want to connect and just basically recreate for your brand. So choose any of these, connect it to your website, your reviews, your whatever, and then just say, hey, can you recreate this ad for my brand? And it will go and do that for you. Um, look, none of this is stuff that you can't do technically inside of Claude or GPT. It's just a lot easier when you have a workflow where I know that like I don't have to go and re-prompt, I don't have to go and re-add context every single time. It's just there for me inside of Poppy.

Jimmy Slagle: And, and the exciting part of this one, as Evan was talking about, MARVIS is our, uh, AI creative strategist name. If you've seen, um, uh, Iron Man, it's a little bit of a spin-off of of Jarvis, but, uh, MARVIS is more for marketing. Uh, and so there's going to be six workflows that we are coming out with, which we are super excited for. We've been working on these for for months, and we are pumped for you guys to go in and actually start to be able to use them. And and seriously, let us know your thoughts, your feedback, anything that you want us to create more or add on to the the great part about MARVIS is it's super easy to go and add more workflows to, uh, uh, its its workflow kit. And we'd love to just continue to work with you guys as much as possible on that. There's some really cool ones. There's one about iterating on your winning ads, so we'll scrape your your winning ads, uh, figure out which one you should be iterating on, and then give you a list of three iterations to make. Um, another one that's just going to look through all your customer reviews and your website to find tons of information like pain points, objections, failed solutions, and more. Um, another one is going to watch your top video ads and tell you exactly how you could turn that video ad into a static ad, whether it's looking at the hook or just another cool scene or or powerful scene within that ad, and many more. And so, excited for you guys to go in and test it out. You guys are literally going to be the first people to go in, um, and and start using these. So, let us know. We want this to be built for you, um, and get as much feedback as possible.

Jimmy Slagle: Okay, so, uh, we we just covered the different AI tools, uh, that you guys can go and add right now. Um, but now we want to talk about one of the most important things, which is becoming an A+ prompter.

Jimmy Slagle: Uh, since everything is around prompting, um, there's really three things that I would say that you should you should be starting with. Number one is just practice. Um, test out new prompts, uh, try to get creative with the prompts, really push AI to the limit of what it can and can't do. Um, and then I'd say the second thing is just learning the best practices on how to prompt.

Jimmy Slagle: If you see the next slide, I'll I'll add in some, some resources that again will be included in that giveaway at the end. If you're a nerd like me and you really like to get into like the more, you'd rather read like a 40-page PDF on how to prompt and and and learn all the different components of it, Google and OpenAI have both come out with a prompt engineering guide, um, on how to work with their models. And we will link both of those in the giveaway for those that tune in at the end. Um, and uh, and then you can nerd out if you want and and do a deep dive.

Alex Cooper: Yeah, that's right, guys. Uh, you can't get rid of me. I will be back tomorrow. Uh, and I'll be back with a new friend, uh, with Dara. And we have a lot to cover in that session tomorrow. We're going to go through best practices, training models, uh, use cases, some real prompts that Dara and I have made, image gen, AI UGC. There is so much packed into that session tomorrow. I think that one is an hour earlier at 1:00 Eastern, if I'm correct. So, yeah, make sure you sign up for that. And, um, that is going to be a very fun one with Dara.

Jimmy Slagle: Okay, so, uh, we're going to touch a little bit on AI workflows. AI workflows are pretty exciting. I'm I'm going to keep this high level, um, just because I know that a lot of people here are are are newer to the AI world.

Jimmy Slagle: Um, and just know, like there's a very common misconception. AI workflows are not the same as AI agents. 99.9% of things that you see today are going to be AI workflows. And so that's what we wanted to really focus on for this quick segment.

Jimmy Slagle: Um, so there's a couple different ways that you can go and start building some AI workflows. Gumloop is is probably my favorite publicly available tool for building Gumloop. I've never written a line of code in my life. Um, and after a few hours of just kind of testing different things out on Gumloop, I was able to get get, uh, a pretty good understanding of how to start to build, uh, an AI workflow. Really what an AI workflow is, just so you guys are all aware, um, is is like this example, is you upload, I know you can't see it, but I can just describe it. You upload your brand's, um, URL and two of your competitors' URLs. From there, it's going to go and scrape the ad libraries of your brand and your two competitors that you you listed. And then it's going to give you an analysis of like, hey, here's some opportunities or gaps in content that no one has really started to focus on yet. And then you can combine that into a report. Let's just say you write it within Notion or Google Docs. And then you can get that report slack, uh, sent over via Slack message or or via email or anything that you'd like. And so it's almost connecting all these different tools together with the ability, uh, to use AI. So if you're familiar with like Zapier, um, they're kind of one of the the originators or even make.com of these workflows. I just like Gumloop, um, because I think it's just built after the AI boom exploded. So it's like an AI-first, uh, workflow builder. Um, that's probably the most important one. It's going to be a little overwhelming, but I again, I think I probably took five, 10 hours of just, I'm I'm going to sit down, I'm going to learn how to do this. Um, and uh, and you start to pick up on things pretty quick.

Jimmy Slagle: So, that was the workflow that I talked about. We'll be giving this one away, uh, to you guys as well. So you guys can just throw in, uh, again, if you're here at the end, you can just, you know, upload your, uh, Facebook URL and then your two competitors' Facebook URL and get that analysis. Um, so, excited for you guys to test out that one.

Jimmy Slagle: Um, the other cool thing about Gumloop is you can connect to Apify. And this is a little more technical, so again, just high level. But if you go on Apify and look at all the different actors, this is this is what is possible. So this is where we've been able to build like a TikTok scraper where you just enter in a search query and an interface that kind of looks like this, very simple. This is all that you guys are going to need to know. You're not going to need to see the back end, but you would just upload like a search query about TikTok, and we would scrape your ad account, um, we'd scrape the top 10, uh, TikTok videos of that, uh, specific search query and then tell you like, hey, based on what we're seeing that's working well on organic TikTok, um, and what you guys have been testing on your ads, um, this is what we would recommend for some new ideas.

Jimmy Slagle: And Apify is essentially how we've been able to do that. So, a couple different Gumloops that or, uh, workflows that we've built in Gumloop, a Face, uh, Facebook ad comment scraper. We finally cracked that code, which was, which was fun. The organic scraper that I talked about, ad library scraper, which I I mentioned as well, on-brand analysis, video script generator. You can really do a lot. It's it's a pretty powerful platform once you start to get to know it. But again, that's a little more advanced. If you haven't done step one and step two yet, that's where you should focus first. This is just for those that are really looking to go, uh, go deep into the AI, um, uh, world.

Jimmy Slagle: And so, beyond just those ones that we're giving away, we're going to give away a couple different more. Um, a copycat agent like Alex had talked about, a video ad quality checker, so you can upload a video ad and we'll check for typos, grammar mistakes, potential copyright or infringement, uh, issues. Um, another one where you can test all the different models. So if you want to see how like ChatGPT compares to Claude, which compares to Google, uh, we'll give you a workflow that does that. Um, and then one that's really cool is, uh, uh, you can upload a video and we'll give you the transcript and exactly what's happening in the video. So this AI model can literally see what's going on in that video and gives you like a scene overview. This is great if you wanted to take like 10 of those to put into like ChatGPT or Claude and say, hey, here are 10 ads that we've, we've, uh, uh, used and that that have worked really well. Um, what are some other, uh, similar ones that, um, uh, we could run or create, uh, that would be similar to those winning ads?

Alex Cooper: Yeah. Um, I want to run through this because I want to be conscious of people's time. Uh, but let's talk about how to put it all together and scale AI across your business. We don't just want the stuff that we talk about today here to be relevant to just you as a strategist or you as the brand owner. We want to help everyone in your organization, and making it as easy as possible and removing the barriers is how you do that.

Alex Cooper: So, this is a few things that we've found, uh, to incentivize AI adoption across the business, uh, in a slightly like fun way. Um, obviously, you know, not being stingy and giving everyone access to, uh, accounts, uh, to Claude and ChatGPT. It is worth way more than $20 a month. So, uh, stop trying to, uh, put everyone on a free plan. There is so much more you can do on the pro plan. Uh, so making sure that everyone's got access to the tools they need. Uh, we do internal hackathons, uh, at Ad-Crate, um, where we try and solve a particular problem. We had one recently, which was really good. We we all jumped on and, uh, um, we had one brand that we were brainstorming for, and everyone, like, we had 30 minutes to see who could make the best, uh, ad in image gen. And we had so many ideas come out of that. Um, so like little hackathons, uh, can be really fun. Having an AI experimentation day, so one day a month where you just give your team the time to go and explore new tools, no expectation of, uh, anyone returning anything, but just say, go and have fun and see what you can find. Uh, bounties for workflow or agent builds is a really interesting one. If there's a particular problem you want it to solve, like, for example, the ad comment scraper, just saying to the team, whoever can solve this, there is a bounty or some kind of reward, uh, for being able to crack the code and incentivize people to, uh, figure their own way out on how to build that. Um, we have a Slack channel for sharing AI resources, uh, and wins, which is a pretty straightforward one. Again, easy access to prompt libraries and context libraries, which is not something that we talked about today, but we will be talking about with Dara tomorrow. And finally, you can actually tie KPIs to, um, uh, specific AI objectives. Uh, now, however you want to do that. I mean, honestly, I think the best way is to give it your current KPIs and and put them into ChatGPT. Uh, but then you can make sure people are properly incentivized, uh, to increase their usage of AI.

Alex Cooper: So that's pretty much, uh, the four-step process or like where we'd start in terms of like transitioning yourself, uh, from a creative strategist into a creative strategist prompt engineer. And I do just want to reiterate like the importance of this, going back to what we said at the top of the call. We do not believe creative strategists will be replaced by AI. You'll be replaced by another creative strategist using AI better than them. And like I said, there is a huge opportunity to get ahead now, uh, while we are still relatively early.

Alex Cooper: So, I said at the beginning we would have, uh, a special announcement. Um, and, uh, we can go through that now.

Alex Cooper: Uh, Jimmy and I want to help you guys become creative strategist prompt engineers. We've actually prepared an eight-week program, uh, to help upskill your team from strategists to creative strategist prompt engineers. And this is the same, like, you know, we've been taught by the same people who have been in the Unilever offices, Virgin offices, teaching them about AI. We were just in the Virgin office last week in Miami, which was pretty fun. Uh, and we're going to teach you guys the same things that we taught them, but specifically applied to the creative strategist role. You're going to get access to a lot of resources, SOPs, prompt libraries, AI trainings, etc. Uh, there is a lot that we are packing into this eight-week live training.

Jimmy Slagle: And just to give you guys a little glimpse of what we think this is going to look like from a very, uh, tangible and and and digestible, uh, format.

Jimmy Slagle: Um, so, uh, you guys will be getting a link here for all the freebies that we just talked about, that copycat Poppy board, all the different Gumloop workflows, the the, uh, deep research prompt, um, and and so much more that we, uh, talked about. And so, um, just use that link and and that should be, um, everything that you need and we'll get that sent over to you shortly.

Alex Cooper: Uh, sweet. Um, well look guys, uh, this is an industry that changes, uh, every single day.

Alex Cooper: So if you do want to keep up with the latest and what we're finding, uh, Jimmy's going to put up, uh, some of the links to our socials here. Uh, you can find Jimmy at Jimmy Slagle, me at Alex Gough Cooper. Um, and, uh, yeah, we tweet, we post on LinkedIn, we sometimes make YouTube videos together, uh, about AI and ad creative. It is an intersection that we both live in, and, uh, we're super excited to have the chance, uh, to speak about with you guys today. If you are a brand spending more than 100k a month on paid social and you're looking for done-for-you ad creative, go to adcrate.co. Uh, we have an incredible team who have achieved a bunch of case studies for eight, eight and nine figure brands. And, um, yeah, just keep an eye on what we're doing at Human Squared. Uh, obviously we have that, uh, that eight-week training program. We've got a few more things on the way as well.

Alex Cooper: Um, thank you everyone, uh, dearly for the chance, uh, to talk about AI with you guys today. We've really enjoyed it. Um, yeah, happy to open up the floor for any questions.

Evan Lee: Round of applause, guys. Everybody in the chat, let's show some love. Come on. I feel like, I feel like the vibes I were getting is everyone was screenshotting like mad and being like, go back a slide. Let me pick up what's going on. So everybody,

Alex Cooper: Oh, sorry, I forgot to say, I don't know if that was in in the deck, but like if you go to that landing page, the adcrate.co/makeads2025, you also get the deck. Um, so if you missed anything, don't you worry, it's in there. You're going to get that too.

Evan Lee: Talk to them, Alex. That's literally the first question I was going to ask that popped up a bunch. So you got it covered.

Alex Cooper: Yeah.

Evan Lee: Uh, everybody, get those Q&As in. I'm prioritizing from top down. There's probably a couple different buckets. But guys, let's kick off. So the first question that I'm going to bring up is probably representative of a bunch of other questions that have also popped up. And it's generally around trust. So how much do you trust AI sources for research? Do you ask it to share URLs of articles? You can see what's going on here, but a couple different questions have popped up like this one. Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. Like one of the things we're going to cover in tomorrow's session is how important it is to become good at prompting. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that way you know like everything that it's finding is actually legit. It it just becomes a little more of a problem when you upload like an Excel sheet or a Google sheet or a large, um, uh, document with a lot of different data. And AI sometimes wants to be lazy and not actually analyze the whole thing. Um, and, uh, and and so there, especially in those scenarios, that's when you have to be, um, uh, just cognizant to start to pick up on like, is this AI generated? Is this not AI generated? Um, and, you know, an easy solution to that is just command F if if you want to double check something. Like if you upload customer reviews and it gives you like 10 really good ones, um, you can just go into the Google, uh, sheet and do command F and see if it pulls up or not. But yeah, you you definitely want to be, uh, cognizant of the fact that AI does hallucinate. That's a problem that a lot of people are trying to solve. Um, and, uh, and, you know, don't just take everything at face value.

Alex Cooper: Yeah. The thing I'd add to that is where you can, uh, try and, like, you know, try and turn it into a text, uh, input for for GPT. Like, for example, if you have a PDF document and it's all text, like it's better to paste that in because AI's got a better chance of like ingesting text than it has of crawling a PDF. It might still do that, but it's just less reliable. Whereas if you've got like, I don't know, if you've got a PDF that's got a lot of images and formatting, then like, you know, you just have to like give it to it and and, uh, and let it do its best. And you can ask questions, you can ask like specific questions about certain parts of the document and see if it's picking things up or not. But like where possible, it is more reliable when it's ingesting text versus the links or, uh, or files.

Evan Lee: Really helpful, guys. Alex, I'm going to throw this next one to you first. So we put up a poll earlier about how many people are using AI in their day-to-day. And it's like a thousand people. It's like, I use it every single day. But there's still this level of like how serious are you taking it? I'll put air quotes up. So I know Jimmy's been living this for a while. I'm curious on your end, like was there any experience or advice you have on someone just going from like passive research to now taking it more seriously to to impact your day-to-day?

Alex Cooper: Yeah. I mean, I I guess honestly, I guess it just comes from, as Jimmy was saying in the prompting section earlier, it comes from educating yourself on like events like these as to how to become a good prompter. I would argue besides copywriting and besides the fundamentals of creative strategy, it is the most important skill that you can learn today. So, honestly, my, like, the switch flipped for me when I, uh, realized, you know, I was looking at the outputs I was getting. And if you just ask ChatGPT for like, you know, write me a headline for my neuropathy supplement, it's not going to come up with good headlines. As Jimmy was saying, it's ChatGPT and Claude and all the all the other LMs, they are not good creative strategists. You have to teach them to be so. So, you know, you're going to start with giving it a task, and then you go, okay, this is not good enough. So you give it some more context. And it's still not good enough. So you give it some more knowledge. And then you give it examples. And then you say like, why these examples are like, uh, good or why they are bad. And again, if it still comes up, like if you do all of that and it still comes up with, um, uh, with examples that are not in line with what you expect, you just take those examples and add those to the bad examples in your prompt. And I know this might seem like a lot of work for a prompt. I'm not just necessarily talking about like day-to-day here. I'm talking about like, you know, I like building things that are going to work for me. Like I like building one workflow or one prompt that I can spend a lot of time up front work like building, and now the rest of my team for the rest of time have a solid prompt that they know is going to give them good outputs. So this stuff does take longer, at least up front to build. But like if you have a workflow where it knows how to generate you good headlines, it knows how to do it on brand, and anyone in your team can access that, uh, instantly by just dropping it into GPT or accessing the workflow, then that's a really powerful thing.

Jimmy Slagle: Yeah, even just adding on to that, like custom GPTs I think are really solid for people of just like getting into this, uh, uh, kind of, um, practice of starting to build some some more like custom versions of ChatGPT. So if you've never built a custom GPT, I'd strongly just recommend starting with that. You can upload your brand guideline documents, you can add instructions on how you want it to talk and kind of what your voice is. The other thing too is like, like don't don't be afraid to ask AI these questions. Like you could upload your brand guideline and essentially say like, hey, I want you to, uh, uh, create a prompt that I could use for, um, like a custom GPT that's just going to create content that adheres to everything within this PDF. And then it's going to do the legwork of having to like write that prompt for you. Um, so I'd say custom GPTs are are a pretty good, uh, starting place for that.

Evan Lee: Huge. And we probably have time for one more question here. I think I've been trying to like work my way through. There's so many questions. So the the audience has really come through. But the one that's popped up is quite tactical, and it's just coming back to Slack. So I've seen so many different questions reference Slack. Um, pick it up first of however it makes sense, but how do we develop these AI strategists within Slack or how does Slack play a role in this? What does it look like tactically?

Jimmy Slagle: It's a good question. Uh, unfortunately, there's nothing like non-technical that you can do right now. You'd need to be building on top of Slack's API. You can have the Gumloop workflows that, uh, what can send a Slack message or read a Slack message or anything there. But in terms of like true agents that you're going to be able to interact with and have a conversation with, um, we're we're not there yet. But again, like, like someone that spends a lot of time thinking about this space, that is where people are going to be building. That's going to be the next big opportunity is to just go where you guys are at already, which is within Slack, and start to build these different research agents within there for Reddit, for TikTok, for, you know, whatever you wanted. Um, so stay tuned on that because I think, I think that is going to be a, uh, uh, advancement that's going to happen over the next, um, uh, three to six months.

Evan Lee: Huge. Okay, guys, what do we want to leave the people with in this last minute? Uh, what makes sense? Alex, kick us off.

Alex Cooper: Yeah, I mean, I I just want to say thank you very much, everyone, for coming out and joining us today. We know everyone's got a thousand things on their plate, so we really appreciate it. And, um, all we want to leave you through with is, um, humansquared.co. Uh, go and sign up to become a creative strategist prompt engineer. We want to help you move into this next era of creative strategy, um, in this live training starting June the 12th. So, thanks for watching. Um, go and check us out, and, uh, we look forward to seeing you over there.

Jimmy Slagle: Yeah, and last thing I would say too, um, I've seen a couple messages of of the automation not getting sent through. Don't worry, if you gave us your email, you will be getting that that, uh, cheat cheat. Um, we'll make sure that's working. So, apologies for the delay.

Alex Cooper: Yeah, it's fine. There might be a couple of minute delay on it because it's a it's a Zap, uh, currently. Uh, so it will be coming through. If it doesn't, I will sweep the email list, uh, later on today, and anyone who doesn't have it, I will send out another copy. And I've got an email going out tomorrow anyway, so I'll just link it all there. Uh, so you will get the giveaway, um, at some point. So if you haven't got it yet, please be with us.

Evan Lee: The freebies are coming. My guys, thank you so, so much. Everyone else, you have been absolutely incredible, filling the chat up. I know they got to go back and read it all now, see what's going on. But guys, appreciate it so much.

Alex Cooper: Yeah.

Evan Lee: Uh, Jimmy, should I throw that one to you first and then toss it to Alex where it makes sense?

Jimmy Slagle: Yeah, yeah, yeah, for sure. Um, so, I'd say the biggest thing is like, especially right away, always make sure to to double check the resources. Like never just take AI for its face value. Um, that's why I love deep research is because it will give you the the specific sources right after it's done talking. Um, and, uh, and and that