How I Make Thousands of High Converting Ad Creatives Every Month

All right, let's be honest. Everybody and their dog is playing with AI just now, but most when it comes to ad creative are using it the wrong way. And you know, turn out creative that looks like a bad stock photo married a generic Canva template. That's why our next speaker is here.

He is going to cut through the noise and show you the right way to use AI to make creative that converts. He's the co-founder of Ad Crate where he's helping brands unlock smarter, faster, higher converting ads with AI powered workflows. And today he's pulling back the curtain and how you can do the same. So if you ever wondered how to turn AI from a shiny new toy into a results driving revenue produced machine, this is your moment.

So, let's drop some robot emojis in the chat and give a huge Creative Con welcome to Alex Cooper. Hello and welcome to my talk for Creative Con 2025. My name is Alex Cooper and today I'm going to be here to talk to you about how to make thousands of high quality ads every single month with the help of AI. Notice how it says the help of AI, not with AI.

Very important distinction. For those of you who don't know me, I am Alex Cooper, co-founder of a company called Adcrate. We make performance creative for eight and nine figure DTOC brands. We have worked with uh 50 plus eight and nine figure brands over the last 3 years.

The likes of Hule, Licorice.com, Harry's just to name a few. And I make content helping people make better ads with the help of AI. Um, I actually recently sp taught a course uh over the summer with over 600 students on how to make ads with the help of AI. Some of you here uh may well have been in that course.

Um, so that's kind of what I'm going to be talking about today and how you can use AI to help you make better ads. I want to start off by talking about an economic concept. Um, and this is the laper curve. And the reason I I'm putting this in here, well, firstly I'll explain what it is.

Uh basically this is used to help calculate the optimum rate of tax uh to collect the most revenue for governments. Um but I think it can be applicable to us here um as users of AI. See what I found uh in amongst training uh hundreds of teams uh on ad creative and AI at this point is a lot of us are here we have used claude we've used GPT and we use them dayto-day we feel like we we like we're pretty decent at them but there's this collective thought that like you know we're not doing as much as we can with AI we want to up our output we want to get better at utilizing AI making it work for us and we feel like it's something that we don't know whether it be a prompt or a process test or a workflow or just a system or a way of thinking about AI that we are not aware of that our competitors are or other businesses are and that leads us to feel there is some knowledge gap um as you can see uh and this is perpetuated by a post like this and you know I'm not digging these guys out I'm really not because I could have found hundreds of posts like this um but it doesn't help when you get people saying that like you know replace your entire creative workflow uh with Nano Banana or I just replace my $200,000 a year uh creative agency with VO. Anyone who's tried these prompts uh or these workflows uh will know uh that if you do do this, you end up on the right hand side of this lapper curve where 100% of the stuff uh well not 100% but like a lot of the stuff that you uh put out just isn't of the quality that you want and you end up scaling volume at the sacrifice of creative quality and all your creative ends up looking like this.

I forgot that part. Um, what I found is that most people actually aren't as far as as far away as they think they are with AI. They just need to learn how to discern between what actually creates tangible business value for your creative strategy and what is just out there for engagement uh on socials. And that's what I'm going to help you try and tackle today.

So, what I'm not going to be showing you is how to replace your 200k a year content team with Nano Banana. uh and whatever other AI tools. Uh if there is a infinite ads machine button out there today, I am not aware of it and you know I spend all day every day thinking about creative and AI. Um so you know I I wish I wish it was out there.

It would make my life way easier but to my knowledge it's not. What I am going to be showing you is how to use AI to enhance your existing workflow. Take what already works and make it better, make it faster, uh make it easier with AI, not completely replace it. uh why people are approaching AI the wrong way.

I'm going be showing you some top performing ads using AI that have spent six figures profitably for our clients and where I think the future of ad creative is and how AI impacts that. So I split this into three parts. Uh let's go through these one by one. Starting with research and ideation.

If you can hear a noise in the background, I apologize. It's my girlfriend's dog uh who's clawing on the door to try and come in. Anyway, I'm going to be starting off by talking about the feed because when Facebook and Instagram first came out with their feed, you saw everyone posting for their friends and family and that was the only thing that was in your feed. Uh it became hard for advertisers to blend in because if it wasn't a friends or family post, then uh it was likely an ad.

Then we started to see the rise of creators uh over the 2010s. the first ever UGC. More DTOC brand start to follow up on socials as well and we s saw the first uh evolution of the feed towards an interestbased feed. the for you page uh was created and over the last few years we've seen the Tik Tokification of the uh feed whereby every piece of content in your feed now whether it's Tik Tok whether it's Instagram whether it's organic whether it's an ad everything is hyper relevant to exactly what you want to see now because the algorithm always wants to serve our ads to the people who are uh most likely to Why the trap that I see a lot of brands fall into is making ads for warm audiences.

Those ads for warm audiences perform better because the al algorithm is hyper relevant to exactly what people want to see. That leads to them making more ads for warm audiences and and then seeing more warm audience ads perform and they keep going through this hyper relevance trap that I call it uh of continuously optimizing ads for warm audiences. Barry Hart actually has a great tweet about this that I can link um at some point as well. And the point is that what's best for your ad account might not always be best for your business.

And ultimately, if you want to scale on meta, you need to be able to get out of this relevance trap and make ads for top of funnel, cold audiences. So, what does this mean for you? Well, first of all, context mass ad account. Think about which ads are you making that are actually bringing in new customers versus scooping up existing demand.

Uh sometimes it could be as obvious as like a bottom of funnel, you know, 40% off uh discount ad. Obviously, that's just scooping up existing demand. Um in in the vast majority of cases, sometimes it's not as obvious as that. Uh but think about which ads are bringing new eyeballs in.

Uh not just iterating on winning ads form audiences like I said. And the reason I'm saying all of this is to break out of this hyper relevance trap, you really need to understand how to market to new customers or how to bring in new customers, how to market to prospects to bring in new customers. Um, and the way that you do that is through good research. So I can tell you what it looks like for us.

Um, I think there are thousands of ways to do research out there and you know a lot of them are six of one, half a dozen of the other. Uh, but for us it looks like this. We have a deep research prompt that we put in the different inputs. So, customer reviews, ad comments, support tickets, surveys, great threads, corora, uh forums, etc.

And the output of this uh big deep research prompt is simply a one sheet that has these seven sections. Personas, stage of awareness, pain points, desire transformations, objection, trigger events, and tone. And what you'll find when you do this thoroughly is that your ad scripts, you know, really do write themselves. You're not writing your headlines anymore.

You're not writing your scripts. You're literally just taking the words and phrases that your customers are using and repurposing them in your messaging. And you're just more so like connecting the dots rather than uh actually trying to think of new ideas uh for your ads. So, here's an example of what one of them looks like for us.

Um again, you can be as as detailed or as uh or as brief as you want in these. We try and make them more detailed. We also don't make these things that are done oneoff. Again, a lot one thing that I see that a lot of brands doing is they just do a lot of research up front when they, you know, when a strategist comes into a role or when they get onto a new account or like whatever it is, they don't do too much research ongoing.

There are can't tell you how many times I've seen in our accounts uh where we've uncovered new use cases or new pain points through, you know, new Reddit posts that have been posted or uh new Tik Toks um that have come up. There is always going to be more information that you can crawl on the internet about your brand and you can uncover new insights about your customers if you keep doing this research uh on an ongoing basis rather than just you know once every so often or or even worse just up front as as a one-off thing. Um so I mean I could cover research honestly the professor's talk but I don't have the time to. Uh let's talk about how to actually turn research into ad ideas.

And this I believe to be the best use case for AI today. Uh cuz when we say like using AI to create ads, I think a lot of people think about you know the VOS, the arc ads, the nano bananas, the image gens, they think about execution. And while that is cool and we're going to get on to that shortly, I actually think the best use case uh for AI is in the research and ideation. And for us, we've had a huge unlock in uh building claw projects that are trained on our data.

So we have claw projects for all sorts of things. UC scripts, podcast scripts, bsl scripts, static headlines, uh script rewriters, iterations, basically any task in the creative process. We have an individual claw project that just does that one thing. rather than just an overall like here's the here's our brand GPT we have individual tasks for each stage of the process and what that means we can load that with specific context about that one task uh and the outputs are significantly better than when we just use a normal claw project.

So here's what each of these uh look like. Um here's our ad headline generator for this brand to do. They're actually a male uh skincare brand. Um so each of them has three pieces of context inside of the claw project.

You have the brand context, the domain context and the top performing examples. Uh the brand context is just simply the uh about the brand which I'm assuming many of you already put into your custom GPS or claw projects. The domain context is where it gets really interesting because that is a context document that we've created on everything we believe to be true about static headlines. For example, it has to be clear and concise.

assume no one knows or cares about the brand. You things like this, we've created a whole document almost like an SOP on how to write perfect static headlines and just static headlines, nothing else. Um, so because that context is relevant and specific to this project, uh the outputs are way better than if we just had like a whole document on how here's how to create every single type of ad that you can create. Uh, and finally, top performing headlines.

Uh that is simply uh the top performing headlines the last 365 days and an explanation as to why we can top performing. Um I wish I had a better better way to show this uh but I don't. Um you just have to go and try it try it yourself honestly. Uh but this is the an example output uh from one of our trained core projects versus like a normal output from obus 4.1.

I know 4.5 is now out but like when I made these slides it was Obus 4.1. Um, so she stopped complaining about my sandpaper face. Boy, I I was I was pumped when this came out. Uh, I has a really good headline for uh two dudes, this brand.

Uh, finally, skincare is for guys who hate skincare. I thought skincare was BS until my wife noticed. You know, I I'm pretty pumped about uh a lot of these headlines that he's coming up with in my train claw project. The cool thing is, as you can see here, it says in my prompt, "Rank these based on which ones do you think have the best chance of performing and why." Um, because I've given it the context of how to write static headlines in my context doc, it actually knows what a good headline is and what a good headline isn't.

See, the problem with Claude and Chat GBT when they're not trained is that they're just trained on thousands and thousands of blog posts and articles from all over the internet, not from good direct response marketers. And that's where you need to come in and inject that context. And that's why on the right hand side, you can see that it's actually ranked the best headlines itself. Of all the 20 headlines it come up with, it surfaced probably surfaced the best five that I would have picked.

uh if not four of the best five that I would have picked versus the one on the left, the normal claude opus, it doesn't know what a good headline is. So, if you ask it to rank them, which by the way is a good prompt that I enjoy, uh then um you're not going to get the top ones shown up first. I don't have time to go through uh the full process for this. If you want to watch me build one of these out um for real, then go to my YouTube, as you can see here, this video I uploaded a few weeks ago, which is my full process for how we build these club projects.

And you can see the props we use, etc., etc., by watching this video. Uh other message of inspiration, um there are a ton on here. I'm a big proponent of the organic feed being the best place uh for ad inspiration, more so than any other ad library. Um, as you guys are probably well aware, your competitors are just copying their competitors who are just copying their competitors and none of us know what work.

So, you know, yes, sometimes we do look at ad libraries. When we do, it's usually via ad spy or foreplay spider cuz I think both of those platforms have their merits in being able to analyze what is actually working at least from afar. But a lot of my best ads have been from Tik Tok, YouTube shorts, or Instagram res inspiration. What we're doing on paid is just trying to emulate what's happening organic anyway.

and sell. Uh so um I'm a big big proponent for getting inspiration from the organic feed. Um but even bigger getting uh inspiration from your customers, I would go there first and then I would go to platforms for ad inspiration. However you get your ad inspiration or however you like decide to uh create your ads, I do think it's valuable to have some form of hypothesis tracking.

Now, we have this uh spreadsheet that we uh use to track all of our concepts. Now, we do this because we deliver it to clients. But even if you aren't an agency and you don't deliver to clients, uh even if you're just a team of uh you know, three people uh with a creative strategy or even if you're a team of one, I would still recommend doing this exercise or some format this exercise. Um I want you to look at column G and column H here.

uh we have two um variables that I think really force us to be intentional about our ads and that is the ICP i.e. who are we talking to and then the hypothesis i.e. why do we think this is going to work? When we roll this out to the team, it really forced our strategy to start thinking about who they're talking to and why they think that this ad is going to work for them.

they were a lot more intentional and it helped them with their ads, especially with the way that Meta is going uh with their um persona testing um being the most important thing uh for your variation iteration. Just having uh these things to be able to track I think is going to be very important uh if nothing else other than to help you make better ads. Um let's talk about execution now. So, uh, this is, you know, stuff that usually goes viral on, uh, on toys from LinkedIn, how do you use AI to produce content that sells, um, I think instead of using AIC, which we'll get on to in a moment, the best use case of AI today still is intentionally unrealistic visuals.

Now, uh, as you can see here, what we want to do is, and this works particularly well if you're a problem based brand, is we want to, uh, create visuals that make the invisible feel visible. We want to create visuals that viscerally display the value propositions of our product. Ideally, viscerally display the pain points of our product. Hype Footwear that you can see on the left here.

They're actually a very good uh brand if you are looking through ad libraries for um ad inspiration. I just love how with this one, if you've got uh this foot pain, um you're going to look at this and you're going to go, "Wow, like that strikes me. That that makes me feel something and will stand out to me in my feed." This one on the right is what I really like um as a as an image yen uh ad. Um, here's some more examples.

Uh, I love this one in the middle. Uh, another nerve pain related one. Again, if I have that problem, that's going to stop me right in my tracks. Um, and again, making the invisible feel visible.

Cool thing is AI can actually come up with these ideas. You don't necessarily have to. You can literally screenshot this slide, put it through GBT and say, "Come up with some potent visual ideas for me uh that display the value props of my product or the pain points of my product that I can turn into an image gen prompt and then get to write the prompt." Uh so AI has actually come up with a bunch of ideas for us that performed uh as image gen or nano banana uh images. Now I get asked at the moment like which is which which AI tools you use the most for image gen.

Uh it is nano banana uh although I think image gen like chatgbt image gen is better at text generation. So if you if you want text directly on the image uh I go with image gen. If you don't I go with nano banana. My preference is to generate a nano banana and then add the text in like a Figma or a canva uh or whatever you use.

Um but you know it's personal preference. I just think none are stronger. Another way that you can uh get creative with image gen is to uh write messages in creative ways. We were doing this before uh image gen was a thing.

Now it's just so much easier with image gen. Um for example uh these uh like these writing on skin like we were doing this before image genen they've done super well. I've had a couple of these spend six figures inside of ad accounts uh where you can just write a message on something that's visually relevant. That's the important thing here.

You can't just write a message on anything. If you're a supplement, you can't just write a message, you know, on the wall. Um or ideally you don't want to write a message on the wall. Uh do something that's visually relevant.

For example, this brand on the left here were a fightear brand and I remember having a conversation with them about an ad that they were making where they um wrote on the octagon. Uh they wrote a message on the canvas of the octagon and uh and had the brace the the knee brace for the fighters next to it. So, how can you create something that's visually relevant to your audience and write in a creative way? Because I've seen these do very very well inside of ad accounts.

The same principles apply to VO3 and now uh Sora and again these slides were right before Sora came out last week. Uh but same thing applies. I do think Sora will change the game in that like like actual AIC uh will be a thing a lot quicker than we think it will because of Sora. I think it's a way ahead of VO in in that sense.

Uh but for now I do think that the intentionally unrealistic visuals are still the best use case. For example, this um one on the left here, the hormones cause spider veins. If I have that problem, like wow, that just I love this this one that we make for hollow socks. Um dandruff one in the middle here.

Uh and on the right, I I I pulled this one in. This is for a popcorn brand of ours. Um I pulled this one in just to show that you can do this. This can apply to you if you're not a problem uh based brand.

Um, this is our popcorn brand or like not our popcorn client. And we've always had an issue trying to explain uh the fact that the kernels are flavor wrapped. Every kernel is flavor wrapped. And that's what makes the popcorn taste great.

Now, if you've had Okap Pop, which is the brand, you know that that is a massive thing and it is a true differentiator as to why this tastes better than anything else. Um, but it's difficult to just explain that in ads to people who have never tried the never tried the popcorn. So, this visual that we created really just I think perfectly displays the uh flavor wrapped uh kernel. Um, and uh and we've used this in a lot of our ads for OPOP since and I just really think it's a great way to again make the invisible feel visible uh with AI.

When it comes to VO, Sora, or honestly any AI gen tool, you're really only bound by your own imagination. That's why I think it's so important for your team to have a central swipe file they can share across each other and just learn like pick up by osmosis the things that people are doing in the world of AI gen. Some of these stuff that we have inside our foreplay board are stupid. Some of them are like actual good ad ideas that we uh that we recreated.

But even if it's just stupid things like I just to just to get our team uh familiar with what is like what is possible cuz anyone can make anything for almost zero cost. It's just you're literally bound by your own mind. Having more ideas and more reference points of what is possible with AI. You know this dog podcast for example, I didn't I didn't even have that idea.

And we have a pet brand that we're now going to recreate this for. This was actually a for for an organic page and this went went and did a few million views and we were like, "Wow, I didn't even know that was possible. Now we're going to make that for one of our pet brands." So, you know, just having a central swipe file that anyone can add to, anyone can see, anyone could share inside the company has been very useful for us as a team. And anyway, like this stuff is is like a very defensible uh thing to learn.

Uh because even if some of the the apps are a little tedious at the moment, I do think that this is going to be a skill set that you will need as an editor, as a designer, as a strategy to get what you want from a text to video prompt is going to be very high in demand. Uh, obviously the idea I think is going to be even more important, but the ability to actually actually get what you want out of the AI is definitely worth practicing even if it's a little tedious up front. I want to touch on AI UGC quickly. Um, honestly, I haven't seen too much of it perform over the last few months.

Uh, I don't think that's because people are kind of tacking on to the fact that AI UGC is a thing. Um, I think it's great from getting you from zero to one if you don't have a lot of creative and you are just getting started. I just think that if AI UGC was better or the same as human creator UGC, then we'd all have AIU at the top of our accounts. And I know that's not the case.

So, you know, yes, we've seen winners from it, uh, for sure. Yes, we still use ARC ads, still use VO uh, for AIC, but I don't think it's the best use case of AI right now. I would spend much more time on if it's AI gen making uh intentionally unrealistic visuals and even more so building out the claw projects uh and custom GBTs. I think you'll get way more banging buck than that when than you will for something like this.

So we use this sporadically uh but it hasn't performed that well for us recently. I will say again my early testing of Sora has been very interesting. uh is quite a way in front of VO uh right now in my opinion. Um the only problem is that they don't get rid of the Sora watermark uh as far as I know yet.

Uh they said that may change in the future, but right now it doesn't uh to protect um people because it's so realistic that people can't tell that it is like AI generated content. Cool. So I made the case for AI. I actually quickly want to make the case against AI cuz here's the thing.

It is so easy to get caught up in all of this stuff. Um especially with all the posts on Twitter and LinkedIn, they all go viral obviously. Um but I just want to put a reminder in here that like all this stuff is cool, but just for some perspective, um AI generated ads are still the minority of ad creative. This is our ad creative output uh all clients over the last year and a half.

And as you can see, AI gen ads are rising. They are rising month- on month pretty much. It's a pretty clear trend, but it's still sub 20% of our total creative. Most of our winners are still human creators.

Uh I don't think that would change in the immediate future, but I don't think it will be that long before it does flip and the majority is AI in my opinion. Just perspective cuz like you know I think a lot of people think that we are you know one of the more AI bullish uh like AI first creative agencies. That is true and even for us like most of our stuff is non AI gen. So, you know, just be careful.

Uh, don't forget the fundamentals because honestly, like that's where most of the ads that we create is at the moment. And I think it's easy to just go, you know, I'm just going to replace our whole team or our whole agency uh stack with um, you know, with Nano Banana, with VR, etc. That's just not the case. Uh, even for us right now.

Um, and focus on the fundamentals because as I said in this tweet, the only people that I've seen make good ads with AI are the people who make good ads without AI. None of this stuff is a silver bullet. If you can't make a good direct response ad with AI, uh AI is not going to save you. Um and that's why it's so important when you're looking for uh people to join your team, they've always got to be direct response first.

They've actually got to be good marketers without AI. It's a lot easier to learn the AI stuff than it is to learn direct response marketing uh as someone who's good at AI. And that is validated by a lot of the most successful brands on paid social. If you look at like a Dr.

Squatch, for example, a lot of their topfunnel ads are literally like a really good uh direct response script, an 11 ads voice over, and solid B-roll selection. That is literally it. They obsess on message over format, and they just focus on the fundamentals. I also think this is an interesting um time to think about, you know, as RP becomes more and more AI and people started to get accustomed to this stuff.

Uh is there a case for creating formats and linear to formats that AI will struggle to replicate or struggle to struggle to replicate as I as I cleverly wrote here uh at least for now. Um, and this is a strategic um, adjustment that we made this year. Uh, and I think it's paid off in the ads we've been able to make cuz a lot of these types of ads have worked in ad accounts for clients because they're just ads that signal that like we're real, we're authentic people behind these ads, not just AI like everyone else has in the feed. So, you know, ugly ads, someone just riffing for 2 minutes uh, to the camera.

AI can't do that now. uh at least not to an extent where it's believable. Uh behind the scenes, these warehouse ads or like employee ads, I'm sure you've all seen on the Tik Tok feed. They all do very very well and actually they're a great ad to make for Q4.

Um so behind the scenes organically, you see authentic podcast conversations or or nonp podcast conversations. Uh not just like, hey, have you tried this product? Yeah, I've tried it. Is it good?

Or yeah, no, I haven't tried. Is it good? But like actually to ideally experts reving and probably my favorite on here, authority figures. Uh it's just so like such a underrated uh way to communicate trust and credibility in a feed where trust and credibility are at their all-time low.

Having someone who can speak authentically about your prospect's pain points and their desired transformations um because they are an expert is something that we've seen work in a lot of accounts. Uh and you know whoever you are whatever industry you're in and you will have some form of authority figure whether it's a doctor or an aesthetician or a food reviewer or whatever find someone in your own industry who's willing to work with you. uh get them to riff. Uh give them some bullet points and um just let them go.

I'm I promise you when you get a good one, they will speak way more authentically than any creator could do uh for uh your brand. When you pair this with white listing as well, this can do very very well. Had a lot of ads that are authority because plus white listing perform very well inside the ad account. Okay, let's quickly touch on analysis before we wrap this up.

Uh, I think Gemini is really good for um conducting performance analysis uh with your ads. Again, it's not going to be to the point where it like replaces your media wise performance teams anytime soon, I don't think, but like it can definitely give you a V1 uh of your performance analysis. For example, we upload all of our top performing ads inside of here and just get it to do the stuff that the performance team would already do. uh like saying what's working, what we should do more of, what we should do less of.

Simple as that. And you know, usually it's not to the point where it can just be handed straight up to the strategy team. But if you can give us a first pass and get us 60% there, 70% there, then like that's still a huge win for us. Um as as you can see, this has saved our performance team literally hours per week uh in not having to do this for all of our clients.

Another really easy way to get a quick win out of Gemini is just literally upload your top performing ads and uh and say come up with headlines. I think this was David Herman's idea. So, shout out to him um originally and then I mean he tweeted about it a few weeks ago but like we've been doing this for a while before that. Here's an example uh of that.

This popcorn is way better than it should be. we turn into this static on the right. And as I said, many of the our top performing statics this year have literally come from doing this. Um, I want to quickly round off by talking about where I think the future of ad creative is.

Uh, I'm assuming if you're watching this, you have a creative team or you're a part of creative team. Uh, or maybe you're not, but most people here will have some form of this. You might not have all of these roles. You might have multiple people in all of these roles.

Um, but this will be the skeleton in some capacity. the way that I see this progressing, I don't know, I don't have a timeline for this. Uh, probably sooner than I think it will be. Uh, whether this is 6 months or 12 months or 18 months or or more than that, I think we're moving towards a world in which creatives become the QC and AI does the execution uh the executing, sorry.

So, we all have research agents, scripting agents, static agents, video agents, media buying agents, and you just take a like a step up to do the QC as a creative strategist and say, "Actually, let's do more of this, less of this. Create this ad, kill this ad, or whatever that be." And AI is going to be the one actually executing for you. Again, this is not today. It's at some point in the future.

I'm just trying to give you an idea for where I think the industry is going. And these agents will actually push ideas and insights to you. For example, it will come and say, you know, morning, Alex. Uh, I hope you had a good weekend.

I've put together your weekly insights over the last seven days. I've analyzed customer reviews, posted surveys, ad comments, support tickets. Here's what I found. It's going to be doing proactive strategy and research for you and then giving you the ideas and telling you or asking you to direct it in which way to go.

So whereas a stretcher today might be focused on writing scripts, briefing designers uh and analyzing performance, uh the paid media specialist of tomorrow may be conducting highlevel reviews on AI generated ads, training models, uh and agents with context as we spoke about earlier and orchestrating the whole team of agents uh to make them better over time and improve the ad performance. So what does this mean? It means it'll be easier to make more ads and it'll be cheaper to make ads. Uh we're already seeing this this year, but notice how it doesn't say like make better ads.

Uh because um well, it's not going to have to make better ads. It's going to have to make more uh and cheaper to make them. I do think this is going to mean that the the floor will increase. The ceiling may well do too, but like the four four creative will increase because AI is going to eliminate a lot of agencies, a lot of strategists, a lot of uh creators um with how good it's going to get.

Now, I think there's still going to be a gap between the best and the rest. Um if we go back to here, we'll think about like, okay, if everyone's got a team of uh agents working for them, like what where is that gap? like what's the difference between the floor and the ceiling? For me, I think it's these four things.

The quality of your direction, so what focus and strategy you're giving it, the quality of your feedback, how good is your taste, the quality of your systems, which aim agents are you deploying? Because they're going to be hundreds and thousands of them. Uh, and the quality of your context, which arguably is the most important one. How well are you training your agents up?

So to conclude, um I think we're going to see the value in a creative strategist flip from the actual skill of creating ads itself to knowing what to create. And that's why my biggest piece of advice to creative strategists is to become a hoarder of top performing ads. uh wherever you see a top performing ad whether it's someone on Twitter or in a call or you know on YouTube or just anyone mentions anything about topic ad save it down to your foreplay board your atria board or whatever and become a hoarder because the more the AI does the executing the more it will be important to have those reference points of things that have worked so you can direct your AI to do the right thing. Um so five key takeaways um to wrap up this talk.

Build core projects and custom GPTs to train on your data. Double down on formats that AI cannot replicate. Upscale your team into new AI gem platforms. Obsess over message not format.

Uh hire and hire people who are great direct marketers first because you know ultimately if you just hire people who are good direct source marketers, they will thrive regardless of what happens in AI. And that's my, you know, underlying uh, you know, kind of conclusion that I want to draw on this. Uh, direct response first, AI second, and you will be good. Um, if you want to follow me, there are my links.

Uh, Cooper, Twitter, YouTube, LinkedIn newsletter. I put out on all of those weekly except YouTube, which I'm very inconsistent at. I need to get better at. So, you know, hopefully there's some stuff there for you.

Um, and my agency is adr.com. If you want to check us out as well, I'm over there. Okay, thanks guys.