Why We DIDN'T Build An Infinite AI Ad Machine (And What We Built Instead) transcript

Okay, I haven't told you this yet, Patrick, but yesterday I was in a meeting with your friend of mine, Kelsey Lyric from 265 Holdings. Multiple time AJF Growth or AJF Podcast guest. What's my podcast called? The Andrew Faris Podcast. That's what it's called. Mhm. Client of ours. And he said to me yesterday, he said, "Did you hire a COO or did you hire a software engineer?" And not really, I mean, hire, I guess, partner with is the better word, right? >> Uh What do you What's I I know the answer that I think.

What's your answer? He He said that He said the same thing in Slack and I said yes. Yeah, right. Yeah, the answer is yes. The answer is he hires cuz you spend like all day vibe coding. I do, yeah. That That's not We talked about it before. That's not the long-term goal that that's what I'm doing all day every day, but, you know, we've obviously built some massive software. We've replaced Frame, Notion, you know, we've built our own time tracking software.

We've, you know, we've replaced like basically seven to eight other softwares in into one. And so, it's been a massive undertaking. And it's funny. Um Uh somebody called it masochistic the other day on Twitter. I was like, "Yeah, that's the right That's the right answer. It's been masochistic." And so, right now, let me answer your question really clearly. Um the answer is yes, I'm both of them right now and I think the reason why it makes so much sense to be doing so much coding right now and software development is is this the software we're building is in the aim is to build better operations.

So, like the CTO side of it supports the COO side of it. So, all the software we're building is meant to fuel our operations and how we do our work. That's the short answer. So, by necessity, because Notion because all the other tools we were using were so limited in what they could do for us, by necessity, it just kind of felt like I've got to go build all these things to do all the things we want to do. When you think about that, do you think that's what you planned to do when when you started?

Cuz I I feel like I have conversations. Yeah. No. What So, By the way, side question. Am I the only reason we even still have Notion? Am I the one slowing us up? Like No. No, no, no, no. Do you even Do you even use it still? Are you >> Barely, but I just like I have some tasks in there still that I need to that I need to port over and I didn't know if like kind of the way that I'm the only reason that we still have like the 15 bucks a month for Asana or whatever we're paying for that right now cuz it's my podcast. >> There's a few other docs in there that are stragglers, but we'll get it over.

All right. What shifted for you? That's what I was going to ask. What shifted for you that made you say, "This is the reason why I'm going to spend all my time rebuilding these tools?" Well, the first thing was the technology got to a place where I could use it because I'm not a software engineer. Um to where I could actually use it and and build something amazing out of it. Um and then the second thing that that happened was I realized, "Oh, I can go do all these things without needing a tech team." Because we actually actually like 6 months ago, not even that long ago, we were talking about how we felt unable to do all the things we wanted to do because we didn't have like a massive budget or you know, we we didn't have anybody technical on our team to go build all this stuff that we want to build, you know, there there are other you know, there are other agencies that are way ahead of us or were at the time on the on the software side because they had either massive internal teams or they paid a lot of money um to technical people to build stuff for them.

So, once I realized, "Oh, I can actually go build all this stuff now that I've wanted for so long." That that's what changed is the realization that I could go do it. Okay, so then how did you decide what to aim that firepower at? Like Vibes. >> [laughter] >> I mean, you know what's funny about that is I would trust that answer. That's part of the reason you like partner with smart people. It is the answer. If you Yeah, yeah, if you were like, "I think I think this will probably do it." Then like trusting somebody with instincts that you go like, "Okay, I trust your instinct." I mean, it's funny cuz you and I have actually very talked We've referenced having a road map and you have a public road map and some of those kinds of things, but >> Yeah.

Finally. But Yeah, but a lot of it is like me just going like, "I I trust you. I trust that you're working on the things that you see the most value for." And you have such a clear vision. You actually do have a really clear vision for this. Anytime I talk to you about it, but but like >> How did you How did you decide? Because cuz there's the the question beneath the question is that like one way of approaching this for a service business right now is that spend every minute creating like an endless ad builder, you know, something like like uh Okay, why don't we just Why don't we just bend over backwards to get the the nano banana Yeah. graphic you know, an endless image ad machine or whatever, you know.

And I know there's people who have done that, right? So, but that's not really where our effort has gone. Obviously, we're doing some AI helped generated production in a lot in significant ways. People first talk about Prometheus, but I don't really see that as the main place that you've put your attention right now. Why not? Um a couple of reasons. One, cuz we have such an amazing creative team and we don't need Like that's just not a need for us.

We're not We don't have a problem in terms of like our ability to produce a lot of ads at scale right now. So, that's just like it's less of a problem for us to solve immediately. Um and so, the flip side of that answer is what did I point the fire hose at? It was It was the pro It was the things that we were having problems with or um we we were bad at and needed to get better at or needed to organize better or just needed to have a better process for.

And so, it's really It was really pain points first. Like what are the biggest pain points that I personally have as a COO and then our team has, you know, trying to do their work. And so, that was kind of group one. Solved a lot of those You know, things as simple as like, "Why aren't people actually accomplishing their to-do's?" You know, "Well, because they're in like 12 different places in Notion." It's like, "I got to go I don't even know where all my to-do's live because there's it's just So, that was a pain point like our pipelines were all our creative pipelines were all convoluted and difficult to keep up with.

Um so, just pain points at first and then like once the pain points have mostly been solved, it's, "Hey, wouldn't it be nice if and wow, we could also do." And so, now we're getting to We've got a great foundation set. That like everybody's inside of the app now basically 24/7. And now it's like, "Oh, wouldn't it be nice if we could also uh do XYZ." So, it started with pain points and then it turned into, "Okay, now that we're here and let me actually Well, there's also a vision.

Like there's this This is is this broader longer-term vision piece, which is and I harp on this a lot. It's get all the context into the OS. So, style guides. >> Tell people what that is though, real fast. What is the OS? You say we're here. What is here? To a spot where everything is in All of our work is inside of this app. So, you know, all of our brand guidelines, all of our concept creation, all of our uh news, announcements, you know, to-do's, rocks, how we make like everything.

Everything that powers how we do, you know, what is a long-form explainer, what is a design still. How do we communicate what that is to an artist? How do we communicate that to the AI that's writing the design skill? So, all of these processes, now that all the data and the information is inside of the app, now we can do really exciting things with it. So, now we can go, "Okay, maybe we do want to make the infinite ad machine." Well, now we've got all the context already available so that we can just go write the code and the AI kind of edge functions so that we can We've We've got, you know, this is what the ad needs to look like because we've got the style guide.

Um so, the big lift was getting all the pieces inside the app. And now we can go figure out what do we want to do to us It's like we got all the Lego pieces in there and now what do we want to build with those Lego pieces? Yeah. The the app is the here also, I would say, right? The idea is like we have instead of having Notion or whatever, we have a we have a custom-built web app called AJF OS. Uh and it Our newest Our newest team member told me today we have to change it.

He said He said He literally started today and he told me that in his meeting with you today, he was like, "We've got to change the name of that. It is such a mouthful. AJF OS." It's like >> [laughter] >> Comes in Comes in in his first meeting of the day is immediately shifting things around, which is like part of why we hired him, right? So, the right kind of >> Yeah, exactly. But I can't wait to come up with a better name.

I know, I know. And And new name for the agency and and everything. So, all in time. Somebody Somebody do this right now. Somebody watching this or listening to this, podcast@ajfgrowth.com, email me or leave a comment. Tell me what we should name this agency cuz AJF Growth Yes. There's a bounty. We'll pay What are we going to pay somebody has this the successful name? I don't know. 500 bucks. Okay, I was going to say a thousand. 750.

We'll meet in the middle. 750. You rename it and we'll send you a cool 750 dollars. There you go. Um But yeah, anyway, leave a comment. Any ideas, then do it. Okay, so, AJF OS is this custom app that has all of this stuff. I think I think it's There's like There's like you you gave the answer for the You saw our needs, right? And so, that you started trying to solve all the problems. But I think like that undersells a little bit of your vision for it.

Like when I when I think about it, I look and I see, you know, one of the things we've done now, right? Is you have all of the clients' assets going in there instead of into like some third-party drive folder or whatever, right? Yep. all of the client the clients brand docs in down not only at the brand level but all the but the product level, right? So, it's like if there's if there's a brand, then you know, you have your brand voice and your brand guidelines and all that kind of stuff.

And if And And of course, if a brand doesn't actually have that stuff built out very well, we can help them with that. Yeah. In fact, that's probably one of the best services we >> That's so true. Cuz >> has it. Yeah. Nobody knows it, you know? >> Yeah. Uh and that's that's not a like I never knew it when I was running my brand. Um it you know, it was vibes. Yeah. >> that's actually another big piece of this is like going, "Okay, we know that this is a challenge for you, and so, you know, let us help you Let us help you with that." Yep.

All the way down to do's and don'ts. Like if there was a product like >> Yep. do use this language, don't use that language. Yeah, I mean, we have one brand, for example, that really looks like it should have an apostrophe in the name. Mhm. But it it doesn't It doesn't have an apostrophe in the name. And so, in all of the brand documents, it is just this really simple thing, which is like a do, write it like this with no apostrophe.

A don't, write it with, you know, whatever. Or or like for supplement brands, like you there's like claims that we tell it make sure you don't say this, you know, whatever. Um but it's funny, those are sounds so small, but like that kind of crap comes up constantly in ad creation. And it's a it's a real sticking point friction point. So, anyways, you've got all that information, plus we have all of our strategic context, right?

So, it's not just the brands files and and all that, but it's also all the way we think about what makes good advertising. So, it's like sort of our stuff that And that stuff all the time we're sort of continuing to update I I'm I actually have it on my list to go do some more writing and thinking about that. Uh in light of Andromeda and some stuff I was reading about Andromeda and going like, "Oh, this probably changes a little bit of the way we think about creative." So, instead of me just calling a meeting and telling people that, which we'll do that, too, I'll also write up documentation that I can give to the machine.

And then that machine can actually help us execute that consistently over time, right? Across all of our brands. So, it's So, it has sort of our input as well. Um and then, once you have all that and the entire creative process, the pipeline being done, now you've got this ability at the same time to track like, "Okay, what kind of uh templates Our template library, of course, is in there as well." >> Yep. Um and which ones are we making for each client?

We can now actually like begin to answer a question that I have heard asked 20,000 times in agency life, which is like, "What kind of ads are we making?" And soon we'll be able to connect that to what kind of ads are working. Like we'll be able to sort of pass that down through naming conventions in, you know, at the level of messages, at the level of talent, at the level of design styles, at the level of storytelling frameworks.

And all of that is possible because because there because of this custom app, where that hosts everything there. Yeah. >> not just vibes, and that is not just solving a couple problems. That's a decision, you know? That's like You at some point, and I didn't tell you to do this. At some point, you looked at what an agency does, and you said, "Here's what we're doing all day. This would make the work better." So, what I want you to press into a little bit is why why did you put your attention there in those ways?

What What What led you to to approaching the the problem of of being the COO of an agency, right? Running operations, to go back to my original question, and saying, "This is how I'm going to solve that problem." Yeah. Uh well, our service at its core is I mean, we use the term operationalized knowledge, but it's information and data, and it's uh the correct organization and presentation I mean, when you get down to marketing at its core, it's the organization and presentation of information to evoke a feeling and and and um commit somebody to buy something, right?

And so, you cannot do that. You can't even begin to do that until you understand all the pieces that are going into that that like us just making a little ad that says, you know, it's it's a design still that has some copy on it, um or even like a socially native still, like it It's funny, like when I see people talking about, "Oh, I made this you know, I vibe coded this thing that goes to somebody's website and then, you know, makes like 100 ads." And it's like, "Yeah, that's that's cool.

It's interesting." But but were those the right ads to make? Right. What Who Who are you selling to? What is the message you're trying to get across? What's the feeling you're trying to evoke? Why are you presenting that information in the way that you're presenting it in the ad? Like all these decisions they can be made by AI, but they can't be made without the proper context and the proper information and, you know, the Legos.

You know, it's great I think it's a great analogy. You know, it's like I've the the agency's job is to assemble the Death Star Lego, you know? And you can't do that without the Legos to begin with, and you certainly can't do it without a plan to understand how that Death Star needs to be assembled. And so, by by creating like this appoint appoint my AI at your website and then I'll make you a bunch of ads, like you've you've not made the Death Star.

You've taken like your bucket of leftover Legos and you smashed them together into like, you know, uh something that like I made when I was 6 years old and, you know, it's kind of a little crappy spaceship or something. You know, it's like I don't know how far that analogy goes, but >> No, yeah, yeah. >> that's why like our work is un- uh understanding assembling and organizing information Yeah. >> Yeah. in in pursuit of a goal, which is evoking emotion, feeling, and desire from from the customer.

From the from the prospect. I just wrapped up a search at AJF Growth for a new member of my team that was done led by my friends at More Staffing. It's a another hire from More Staffing for my team. More Staffing, as you know, is an agency that connects you a recruiting firm that connects you to incredible [music] talent in the Philippines. And I'm just so thrilled with the candidate that we ended up with. The whole thing was incredibly easy.

I sent a job description over to More Staffing. They helped me think about whether or not that was the the right job description for the position, clarified that, did all the pre-interviews, had a great pipeline. In the end, recommended two candidates for the position. We liked both of them so much, it was actually a really hard decision uh because they were both so good. And so, we picked one. We would have hired either of them actually in most cases cuz they both they both were really so good.

And all it required from us was one interview with each candidate. So, instead of a whole drawn-out hiring process, they did took all that work off our plate, gave us vetted people who were really good, and did that. More Staffing really understands how to connect great Filipino talent to great [music] US companies. That's because they did it themselves before they were ever a uh hiring firm. They were [music] D2C e-commerce brand builders, Greg Carey and his team, who staffed their team of Filipinos.

They understand the challenges and any cultural issues, all those kinds of things. They just do a great job. If you're looking to make your OPEX dollars go farther in your business and get incredible talent, create operational leverage in your business, you should talk to More Staffing about getting great talent from the Philippines >> [music] >> to your team. There's a whole a whole bunch more stuff I could say about it, but the big thing I'll say is they've been sponsoring this podcast for years.

They've been staffing my teams for years, and I couldn't be a bigger fan of what they're doing. Go check it out. If you've got any positions open, even manager, director, etc. level positions, go to More and see if they can help you fill it. morestaffing.co/ajf. Tell them I sent you. [music] If you have a hire with More Staffing that does not last in your company for a year, they will replace that person for free. No additional headhunting fee or anything like that.

So, uh >> [music] >> really awesome. Most Most headhunters are just way shorter than a year for that, but they they they really want to set you up [music] for success. So, morestaffing.co/ajf. Go check it out today. Link's in the show notes. Yeah. And so, it's very easy to look at an any individual ad and say like and say something about that ad. What's really hard is actually replicating that at scale across clients for different with different styles and different voices while maintaining the strategic things that you think are important for the for the particular work of Meta Ads advertising, right?

Like There are There are There are things that are true about that pursuit that are true for every brand. It's sort of Yeah, in some ways it's funny. It comes back to this thing that I kick against all the time, which is this idea that like every brand is different, you know? People will say that kind of thing. There's no There's no way to uh There's no way to uh like like test it for yourself. Every brand is different.

You know, the idea of a playbook doesn't work, whatever. I think that's just wrong. And yet, of course, there's a grain of truth in it, which is that every brand does have distinct elements of the brand that need to be communicated differently. And that that mix of things, the fact that every brand is different and at the same time, there are strategic things that are true across all brands and and that are important for the for the work of Meta Ads advertising more broadly, the tension between those things is exactly the problem of agency work.

It's extremely hard uh to get those things right. And like one of the solves for that is trying to find ways to consistently marry both sides of that information spectrum or whatever, both of those things into one repeatable process across brands, so that you can actually bring the best of the general principles to the individual output. And that means having a combination of the general principles and the and and the individual assets or or whatever it is, right? >> Yeah.

So, yeah, I I I you look like you're going to say something about that. Yeah. Well, I I'm certain that It's funny, there's like you know, there's the age-old debate of, you know, agency versus building house. And there's there's good there's good uh arguments on both sides of it. But I'm I'm so certain that the way that we think about and um organize and approach the the the work of understanding a brand, uh its marketing moments, its products, its customer, and all the pieces that go into really good advertising.

I'm certain that we do it uh better than most brands. Uh and I can't say that for sure because I don't know what's going on in most brands, but I used to run an an eight-figure brand. I know a lot of guys and gals that run eight-figure, seven-figure brands. None of them have ever thought as hard about, including myself and my old brand owner, have never thought as hard about doing what we're doing in terms of I don't want to I want to I don't want to oversimplify it and just say it's organizing information because that's not it, but it's that nobody has thought as hard about what we do um as much as as as we do.

So. Yeah. Yeah, I think that's one kind of And and getting to do it across so many brands >> Yeah. and for so many different types of brands has allowed That's been one of the probably one of the biggest challenges is like how do we create systems and processes and playbooks and approaches to advertising that work well across everything from a supplement client to a jewelry client to a you know personal safety alarm, etc.

Like how That was a big piece of trying to understand how to fit all these pieces together is how do you do it in a way that is flexible across any sort of brand you could ever come across. And I think we've done a really good job at at developing that. But there's another layer to it, too, right? Which is not only that element, but the just the more sort of baseline operations side of things, which is like how do you help people This is like the you know, the creative supply chain concept.

Like how do you if in that if creative is a big part of what we do and we think we can really make a difference by providing more and better and more diverse creative across brands all the time with our clients. And I do think it's important and it's a and it's like a critical part of how we operate. If that's true then uh then how do you actually make that process work well? I would say that seems to me to be the other big investment that you have made.

And I'm curious to hear you talk more about why you've done that instead of just make the infinite ad machine. Like why why I don't know. Like why invest in sort of like creative analytics not at the level of like performance but at the level of like the supply chain, right? Like how many ads are we producing? What kind of ads are we producing? How how how are different editors doing? Rating it by complexity score, FPAR, which I've talked about, right?

First pass approval rating. So how many ads are making it through the approval process on the first pass every time? Like All All of these things that are like internal. The clients basically never going to see a lot of this stuff. They'll see some of it, but a lot of it is just like organizing our work that way to make it to do that. That's like been another part of this investment. Um so yeah, anyway, talk about that.

Yeah, I uh I'm going to try to pick up that thread. Let me know if I'm not, but like it >> it really just comes back to So I'm an engineer by background. That's That's how my brain works. Um and engineers work from first principles, right? And the first principles for me, I can't go make a bunch of AI ads without first understanding what it is that that needs to all of the context that needs to go into making that ad.

So it's uh it's a it's Yeah, I don't know how to say it other than it's first principles and in order to assemble the right work in order to give our artists the right direction and eventually AI the right direction or whatever we're going to Let's Let's take another example, landing pages, right? That's a big thing that is going to be a new service that we're going to be offering soon is landing pages. Okay. Yeah. Well, it's one thing to go into Figma and design a landing page.

That's great. Any artist any designer can go do that. It's another thing to figure out how to deploy it to Shopify to tie in the right offers into the back end to make sure that all your pixels are set up, you know, on and on and on. It's >> Yeah. Like And this is actually something that really pisses me off about AI Twitter X is like you're always seeing people post, "Hey, you know, I made I made this app that you know." In fact, there was this guy I kind of I kind of ragged on him a little bit the other day.

He was like, "I made this infinite AI ad machine that you can go make a hundred ads, you know, with this Claude code that I set up in 15 minutes." And uh that's nice. It looks pretty on the front end, but he did no actual, you know, back end work of powering the databases that are required to to power that. It's like It's one thing to show a pretty front end. Like anybody can go into Lovable and put together a nice pretty dashboard that looks like it's doing something for you.

But it's actually not because there's no back end databases. There's no back end um there's no engineering behind the the the actual tool that that person has made. I don't know if I answered your question, but that's what comes to mind. >> the problem with that? Like I mean, what what what's the problem with the lack of engineering there? Um Why isn't it enough to Why is it not enough to make the ad machine? Well, in this particular example, the guy hadn't actually made an ad machine.

He just made something pretty that he could turn into a TikTok that looked like it was an infinite ad machine. Uh-huh. >> Um you know, it's another thing to hook up all the APIs and you know, generate the AI edge functions on the back end that tell tell the machine how like I don't know how to describe this. Um I guess my my my uh my encouragement is don't trust everything you see on Twitter that, you know, I did this I did this in 15 minutes.

It was really easy. Um Yeah, it is really Okay, back to the landing page. It is really easy to put together a landing page for a client in uh what's Google's new I think it's Google Studio or something. I forget what it's called. It's It's super easy. Just one prompt. It's going to look beautiful. Okay, that's nice. Now what are you going to do with it? Where are you going to host it? How are you going to get it on Shopify?

You know, how are you going to integrate it with your offers? How are you going to integrate it with all your variants? Like if you're selling a shoe in 12 different It's like it just goes like the the um the rabbit hole goes deep after the fancy landing page. There are all sorts of decisions and pieces of what I call engineering beneath the surface. Not like in a technical sense, but just like first principles. Like um Yeah.

I don't know if I don't know if that's making sense. Well, and and I think one of the things that that that points to is that like a lot of what I see on Twitter or elsewhere is very short-term. It's like It's like It's like I did a thing once that is like a useful thing. And that's fine. But one of the things that I think about when I look at what we're trying to accomplish is like you know, that that the thing everybody says about AI, right?

Is the breakneck pace of change. And so Mhm. And so like building something that is sort of able to sustain something that we we look at and say, "This will allow us actually to uh to um This will allow us to uh constantly add different tools into that are coming out all the time onto a base layer of how the work gets done that we can then improve the processes over time, right? If you just like stretch this game out three to five years you then building an infinite ad machine today may have some usefulness for a short amount of time or something like that.

But I And and you know, I think there are some people who are really good at being like very fast movers and making little tools on and taking advantage of moments, but it's just sort of not what I'm the most interested in. What I'm What I'm really interested in is like how do you build something that allows you to create value over a very long time in a deeper partnership and a deeper >> Yeah. like more durable kind of business that is like you can really build things on top of it and allow it and allow it to to be a a sort of long-lasting machine.

And I don't see great evidence yet that that's not still going to be the most valuable thing in all of this. And so Mhm. uh And so So for me like any individual bespoke thing is fine. But of course, what also everybody knows about like your you know, in 15 minutes I built an infinite ad machine thing uh is that like actually of those hundred ads or whatever, like four will be usable. And the other like 96 will be like semi-usable, but then a designer will have to come and like touch them up or whatever. >> still there's still a human piece needed. >> Yeah.

And and and while it's the case that, you know, there's there's some counterarguments to that that I can see, but but um but I I just think instead it's like how do you build it so that every the whole process of how we do the work we do. And we're talking a lot about the creative process, but I think this is also going to apply with the way we do forecasting and with the way we do media buying and some of those things as well.

The more clarity we have to each part of that process in the actual the clarity in terms of how the work gets done, the more that we can use tools as they become available and as they're available right now to make that work better and faster and then to use humans to um to do the most human most interesting parts of it, you know? And it's sort of like >> Yeah, so anyway, that that's like that's part of what I look at.

And it took me a little while to get there. First I was like, to be honest with you, like a little frustrated going like "Why why aren't we making" I felt this like anxiety about like why don't we make the AJF growth infinite ad machine? Like shouldn't we be doing that right now? You know, I see people doing it. It seems to work." And now I'm like really fully bought in where it's like I don't know. We'll probably add like AI image gen ads as part of the process at some point here.

Right. There's already a V1 in in the system. It's fine. Um Yeah. But there's much there's a much more interesting thing here, which is like, "Okay, well how do we how do we actually instead do something so that when we build that infinite ad machine, it's durable? Like it has access to all of the assets in the asset library from the client. It has access to all of the documentation. It has access to like, do you understand?

Like, it can do all of those things and you just keep building on top of that, but there's this foundational layer that everything starts with. Well, there's also the team side of it that I think maybe, and this is not a criticism, but maybe you weren't thinking about when you were frustrated, which is it's one thing to build the software, it's another thing to integrate it into the team. And there's a couple there's a couple you know, here's a couple examples.

What happens when you send that ad to a client and they don't like it? Yeah. Who's Who's Who's going to re-edit it? Like, is it Is the AI going to re-edit it? No, you need a human to go back to the AI Like, there's always going to be, and this will change over time, but there will always be a human in the loop and somebody needing to operate the AI to to to accomplish the goals that you you eventually want. And that may be you know, some agencies out there, that may be one person doing it for 12 brands, maybe.

Um, for us, that's not our approach. Um, I think I I guess what I'm trying to say is even if I built the most amazing phenomenal ad generation machine ever that nobody ever has to do any work ever but like log in once a day and click one button, there's still going to be a human piece needed where a where a client's going to go "Oh, can you swap out that image for this one?" Yeah. "Oh, I like the copy there, but can you change it to that?" Oh, you know.

Hey, and it's always in Slack. It's always like, "Oh, drop that, you know." It's like always And that's not me complaining, but there's always like It's funny cuz I see our our client Slack channels all the time. I'm I like checking in on them and I'm like, "Yep, AI can't do that. AI's not going to help there, you know." That request, you're not going to have an agent do that request and, you know, um so yeah, my my point is like, well, it's one thing to make the tool, it's another thing to like integrate it into our processes and make sure it works in a way that that pleases the client.

And right now, we have a great process that the client loves. So, I don't want to go That's my I'm not interested in breaking that right now. Yeah. Um, I'd rather go make a landing page service rather than replace replace people that are already doing great work for us. The entirety of the creative process that Patrick and I are talking about in this episode is staffed by designers and video editors from the Philippines, creative strategists from the Philippines, through my friends at Behind the Scenes Studio.

Behind the Scenes Studio is a sister company, maybe, I don't know, daughter company, something like that, of More Staffing, who's also sponsored this episode and of this podcast. And [music] they deliver a sort of agency-style service for Meta Ads creative that is is a creative house that you can use either in-house as an extension of your own team so that you can be if you're a brand that's building your creative process in-house and [music] you want design and edit help for that sort of thing, Behind the Scenes Studio can help you there.

Or if you're an agency or a freelancer and you need creative help on your team, BTS can help you there as well. And they just do incredible work. I'm a huge fan of all of them. I had their CEO, AC, at my house for dinner last week. She's become a critical partner in my businesses. She's thinking all the time about things that we're doing at AJF Growth and how they can apply those across their clients as well. If you're looking for design and edit creative resources to make a bunch of Meta Ads for you, integrating AI, all of those things, and that's really the extent of the service you're looking for, Behind the Scenes Studio is a great place for you to go look.

You should get on a call with them and see if they can make ads for your business that are good. They have made now over 100,000 ads in their life [music] span as an agency. They just do really, really good work and I'm a huge fan of theirs. So, go to btsstudio.co, btsstudio.co, get on a call with AC or with somebody from her team and and see if they're the right fit for you. If you need ads for your Meta Ads and elsewhere, get on a call with them and see if they'd be a fit.

They're staffing our team. I'm a huge fan. They should staff yours. And critically, like, there's one way of hearing what you're saying there, and I know this instinct is going to come up for some listeners or or viewers of this, which is you're going to go those clients are so annoying that they that they want those little changes, you know. And that's I think sometimes Well, sometimes it's true. It but it's not true of any of my current clients.

Uh That's right. The The most times that's exactly what the client should do. The client Yes. It The client should be trying to build something durable, too. If you want to build a great e-commerce business, I Yeah. I had lunch yesterday with uh with Bear Handlin, one of my very favorite operators in the e-commerce space, Born Primitive. And that brand is such a perfect example of like what it looks like to build something really durable that has a real point of view that has that has has real brand loyalty from its customers.

And you just can see it clearly, you know? And they have built that over 12 years of just like Yep. you know, starting Like a lot of brands started off like showing up with a booth at, you know, a a local fitness event or whatever and like selling some shirts or something. From there, all the way to all the way to like building a really good mid-eight-figure business today, you know, um with that is doing really, really cool stuff.

Partnering with huge everything, whatever. And the thing is, that what is required to do that, in part, among other things, is to have a vision of your brand that is going to create some kind of real value in the world for some kind of customer. And so So, the idea that they that like you're just going to do that with AI is like No, like, you want you want to build something awesome. You have to build something awesome.

And it's only going to be more important that you build something awesome because the barrier to entry is going to get lower. So, so it becomes more incumbent on brands to care a freaking bunch about about like what they say publicly and what works publicly and to hold agencies to building great work because of all that, to say nothing of the fact that I think increasingly you're going to have people sort of having to declare what's AI and what's not and things people suspect things, whatever else, right?

Yep. Um, so to me, like, that's what you want to build for is like durability and the long term. Um, and yeah, in fact, uh I I talked to somebody to somebody else who's an investor who said, you know, he's he and his uh partners are looking not for um not for like uh really high returns on their investment. What they're looking for is consistent returns over a long period of time. So, he said that we would rather instead of it's we would rather have instead of getting a 40% return one year and a 10% or 5% the next year or whatever, like, we would we would rather get 20% every year for a very long time.

That's the way we're looking at like their investment. And I think there's something where brands ought to be thinking somewhat that way, too, where it's like my goal is not to get like to a hundred million dollars tomorrow. My goal is to build something consistently over a long period of time that's actually great, that resonates. Uh By far, the best brands that I see in D2C are the brands that think that way. And a few brands are able to think that way and move extremely fast at the same time, and that's great.

But what I think is more interesting to me is the idea of the brand that goes like, "We're going to grow this much every year for a very, very long time, and we're going to build something really, really awesome. And whether or not we're the ones who own the brand in the very, very long term when it gets here or we've sold it somewhere along the way or whatever, that's the way we're thinking." So, if you think that way as a brand, then your agency and the people who are doing the work for you need to be thinking that way, also, which is like, "How do we go really deep getting into everything the brand is doing, creating great work?" And that just requires a different kind of thing than I built in the ad machine in 15 minutes, you know?

Preach. >> [laughter] >> Brands should do awesome stuff. That's my point. >> Yeah. And it's so it sounds so silly, but like it's it actually it actually is different than like what what I think a lot of brands are out to do. They see AI I mean, it really does It's like a get-rich-quick scheme to people. And it's like, "What You know what everybody tells you about get-rich-quick schemes? Don't do them. Don't do them." That's what everybody tells you about them, you know?

It's like, "Don't do that." Instead, keep building for the long term, you know? Yeah. I like to think of AI as like or when I think about the software I'm building, it's how, and I say this a lot, but how do I automate and improve all the things that that is not the real work, so that our talented people can do the real work. So, even I'll even go so far to say as, yeah, let's go automate socially native stills as much as possible.

Right. So that my creative strategist can think really hard about big, you know, we we've called them moonshots in the past. Like, what are these like really exciting long-form videos, you know, founder story, the kind of things that take real deep thought and time and effort and like I want my creative strategist actually spending all her time on that rather than >> Yes. where does the sticky note go in the socially native still?

Um, you know, I was thinking about this last night. Like, how do we start? This is a conversation you and I need to have. How do we start calling our team to higher greater things? Going, "Don't worry about that anymore. That's important. It's but it's getting done. And um a better use of your time Like, the way I think about it is identify the things that robots can do and identify the things they can't do and go spend all your time on the things they can't do really well right now.

And so, one of the things is, well, they can't service the client super well right now. They can't develop trust with the client. They can't be deep in the Slack channel going, "Whatever you need, I got it for you. You can trust that you asked me to do something, it's going to be done. You can trust that like when you when you say this really nuanced thing about this creative, I'm going to remember it because AI ain't going to remember that." So, I don't know.

That's what I want to keep calling our team to and like coaching our team on is like for forget about like let go of the things that like you know, like we're getting to a point where our concept briefs are being like we need to just stop doing any human interventions, you know, yeah, whatsoever. Like quit wasting our time tweaking little bits of copy that don't matter. Yeah. Um, because we're probably actually making it worse in a lot of cases.

And just just spend more time on the things that that humans are a lot better at than than AI is. We have a We have an ad right now that's become the top spender for our client. It's I'm just I'm so excited about it because it's one of those clients where the brand has existed for a long time and the top spenders in the ad account have been the top spenders for a really long time. Uh, like for years. And there's like no nobody who's worked on the brand last couple years seems to have been able to beat those top spenders.

Well, we beat it in 2 months, okay? Um, and I So, first of all, that feels great. But the ads that are winning are really interesting because it's a very simple template, uh, but it works perfectly for the brand in this category. I desperately wish I could show it. Um, and it and what is under the surface of the ad if you were just looking at it is like really thoughtful things happening at like a bunch of different levels.

One of them is it's a template that the visual template of the ad is a is a template that I identified from the organic feed a long time ago, okay? So, there's my claim to it. Like this there's my part in the process. I look at this sort of visual arrangement format of an ad and I thought this it was actually an organic post. I thought this would really work. And then I tweaked it tweaked it a little for the particular category this brand is in and and so I recorded the template.

You built the software that makes it so that that template is easily accessible for a graphic designer. So, that when I when I have vision for how this ad was going to work for this client, I could say to the graphic designer, make it look like this. The this in that equation uh, is easy easily accessible. It didn't require me to slack somebody something and say, "Hey, can you make it look like this ad?" You know, like what would have what would have been the case a year ago.

I would have had to say, "Let's make a couple ads like this one." and send them a slack and say, "Here's the format." and I wouldn't have carefully detailed what I meant by make it look like this, what parts of it I cared about keeping, what parts of it I didn't care about keeping, all that stuff. I just would have sent them an example. Um, now we have an an a thing identified. Here's what makes this this a compelling visual format to convey the message that we care about, okay?

Um, it also comes down to messaging that I worked on with the client with the creative strategist on this account. Shout out, Cara. And uh, and uh, and and kind of looked at together based on my analysis of the ad account. Then that messaging made it through our pipeline along with multiple ad formats or with multiple variations of the ad uh, and critically the graphic designer picked it up and the graphic designer didn't have the assets that we would need to make the ad format that I wanted.

Like they they just didn't exist in the client's ad asset library. So, you know what she did? She took the assets that were there uh, after the after Cara uh, wrote the copy that was beautiful, started with Prometheus, then tweaked it to make it fit a little bit better because the niche she needed a human there to get exactly what she wanted. Her improvement on Prometheus was real. It is way better than what Prometheus spat out at first because she saw what the need was there.

Uh, okay. And uh, it's like it's like one line of really really good copy that feels really elevated. And then with that copy in hand, the graphic designer, shout out Pauline, uh, Pauline turned around and took the assets that were existing and said, "Oh, those don't work for the visual format that Andrew and Cara wanted here. Okay, great. Uh, I need to go and use AI to generate visuals that now work here." And the visuals she generated are so beautiful, Patrick.

I haven't even showed you this yet, but they're so beautiful. I actually was on the call and didn't realize that they were not original assets to the create to to the brand. I said something to the brand yesterday. I said, "Man, part of the reason this works is because the photography is so good." And and the brand was like, "That's not our photography. Your your your designer >> [laughter] >> your designer generated that with AI." I was like, "I had no idea that was the case because it's so perfectly done and so good." >> let me stop you.

What like 8 months ago you were like, "I hate AI images. They're never going to look real." >> Well. >> [laughter] >> Well, but but the thing is this is actually part of the point. Is there something more interesting to me than the infinite ad generation machine here? What was it What was interesting about this to me was the instead of saying the robot replaces the human, what happens if you give the human the power of of CGI, you know, in their pocket?

And you say and you say, "Hey, you now have a like the example I was using right now. You have a way more powerful hammer to swing than you used to. You were just limited by how big your hammer was before, but now you have a giant hammer and giant muscles to swing it, okay?" Um, and so like what happens if you do that? Well, you get a really creative person like Pauline who's really good and you give her that tool and you get a really creative person like Cara who's a really good writer and you you know, whatever.

And yeah. And what happens is you get the top spending ad in the ad account that nobody's been able to beat the top spending ad for years, you know? Um, and it's because and at every part of that process there was something that got touched or built that made that possible and it can all get way better still, you know? So, it's like really compelling to me what you're saying, which is this idea of like once you've laid that foundation visually, then you can go, "Oh, we can go and say there's a kind of messaging that works for this client.

How do we replicate that? How do we lean into that? How do we make more versions of that? That also fits with what Andromeda is right now and what what is good for Meta and there's a visual style. It's from the organic feed. There's all of these elements that are all coming together and the task at hand of the agency is to keep do is to keep making every part of that process work better and smoother so that the designers do cooler design, more interesting stuff.

It elevates the brand. All those things work together. You know what's what's really interesting about that story you just told is, you know, 10 minutes ago we were talking about, "Oh, we don't have the fancy AI ad generation machine. We must be behind." Like that's the feel on Twitter. But what what you just the story you just said was the the only part of it that wasn't AI assisted was the part of you, you know, identifying the ad.

But you identified the ad. I used my AI vibe coded software to organize that ad properly, present it to the team properly. Cara used our AI copywriter to write the initial copy, but then she had the human piece added on to it. Pauline used AI generate So, like >> She also used it to generate the concepts and to like and so sort of like like and to So, it's not that we're not ask. Yeah. It's not that we're not like on the bleeding We are we are very much on the bleeding edge of AI capabilities.

It's just not the sexy X post AI capabilities. Like Like all of that was an extremely uh, tech leveraged and very clearly thought out >> Yes. uh, operation And >> that ended in the the final ad. Which Which when you if you showed that ad, people would probably be like Oh, yeah. Yeah, they wouldn't care. They wouldn't care. Well, but see it well, the the other thing I was going to say, I'm sorry. The the I'm too excited about this.

The other thing that's going to happen in this process is that we're going to go to an in-person with that client like in a month, uh, month and a half. And one of the implications we're going to have is that like, "Hey, there's a there's a there's a visual style here that feels more elevated relative to your price point." So, it's going to make the product look more premium than it is, basically, right? Or or I mean I don't know.

That's that's It is it's a well-done product. It's not an expensive product. So, It feels expensive when you look at the ad, but then the price point >> That's exactly right. So, it's going to So, it's going to make it feel like a really good value to the customer. So, there's there's an which is an act of communication. I want to be clear about that. That is messaging, okay? Um, so so when that that happens and then we're going to be able to go to the client and say and look at this copy.

It's not like buy this right now, it's good. It's very moody and brandy in a way that really fits with this brand and also speaks to a very specific customer. And we're going to be able to take that and say, "Do you see this is resonating in the ad account? This fits with a lot of what this brand has been about historically. How do we now go in our product developments and our influencer selection and our organic social and all these things and incorporate those same elements so that we create durability in this brand so that the brand take continues to take on and live out this voice and do things that a brand that that resonates in these ways with this customer would actually care about?

How do we show up for our customers in deeper ways across everything we do that keeps reinforcing this cuz this is what's really resonating and and we've been able to show that really effectively. And then how do we take that to the rest of our creative as well? So, it can it can it could because it just it takes deep work to do good brand work. It takes real thinking and that's why there's this layer of it that's beyond the the week in week out blocking and tackling of creative work that actually like kind of goes out to this other layer and say, "Here's how we resonate." You know, that's the conversation I want to have alongside like, "Okay, and what's happening financially for the brand and how are our targets and our costs coming in where we thought and all that stuff, too?" Which is really important, also, you know?

So, to me there's like there's all of these kind of layers of that that that go beyond, you know, like I said, the get-rich-quick scheme of of infinite ad AI, you know? Anything else to add? No. I got more excited than I expected to, Patrick. >> No, that's great, man. Bring it. Bring the passion. It's fun. It's fun to see. It's fun to see great work get delivered in ways that's like Yeah, that would have been a $10,000 $20,000 photography shoot across all your product collections before and now we can just like bring that to people in a way that's really cool.

And Yeah, we should we should start calling these podcasts our celebration episodes. We don't we don't take much time to sit back and just go, "This is amazing, the work that's going on." And uh, anyways, yeah, it's fun to celebrate. Yeah. All right, I'm going to go post that out on Slack right now so that everybody can see it and I can replicate the excitement I just had. So, I got to go. I got I got I got to go do that.

It was all fun, man, as always. Okay. See you. See you. Thanks for watching, for listening. As I've got a couple of great episodes coming up. You do not want to miss my episode with Andrew Youdarian. Going to be a packed episode. It It should come out next week at the time of releasing this about his report showing as the e-commerce field trends report where he's got a whole bunch of data from 300 brands [music] that he has looked at, surveyed, and seen what is the real state of DTC brands right now.

We went through all of that data. There's really surprising stuff in there. How's AI adoption going? How many brands are currently on Amazon? And how good is that a part of as part of their business? How bad is that as part of their business? Is that rising? Is that falling? Um all kinds of stuff like that. Are Meta ads [music] brands doing better or worse? Are tariffs making things uh worse and are they killing brands?

All those kinds of questions. We get into all of it. If you really want to understand where you are, subscribe to this podcast right now so you don't miss that episode with Andrew. It's really, really good. Also got some stuff on media buying basics and not really basics even, but just things that I think really matter that people still get wrong coming up. Solo episode. I've got an episode on retention coming soon. A whole bunch of really good stuff.

Subscribe. Don't miss it. Leave a comment on this episode. I engage with all of them. I read all of them. I comment on all of them. Actually one of the best ways to get a hold of me. So leave a comment on this episode. Tell me what you think. If you're interested in working with AJF Growth at some point, we're really tight on space, maybe totally closed at this point, but you should get a conversation going with me now for as we keep for when we keep growing to see if we're the right fit.

Even if now's not the right time, shoot me an email. podcast@ajfgrowth.com. Or go to ajfgrowth.com, fill out the intake form, tell me a little bit about your business and we can get that going from there. Thanks so much to More Staffing and Behind the Scenes Studio for sponsoring this episode. Go check both of them out. The links are in the show notes. Huge fan of them and everything going [music] on with adding great Filipino talent to your team in the US.

Thanks so much for watching, for listening. I'll see you next time.