Why Are Early AI Adopters' Businesses Performing So Badly? (With Patrick Coddou) transcript

Then you need to try our laundry soap. It's made with plant-based ingredients that are biodegradable and safe for your grey water system. >> [music] >> Okay, Patrick, that that that hook that video was so good. [laughter] It's so good. Uh and that that you want to you want to say what you found there and how you sent that to me? Uh so I >> Yeah, so that's a new announcement from Higgsfield. I don't know how new it'll be when this publishes, but it's this kind of great way to easily make hooks for videos, AI hooks for videos, which by the way, we talked about like 6 months ago on one of our podcasts.

So we wanted to start doing more of that and uh for the audience, we have not followed through on that. So I saw that tweet and I was like, this is going to be a great way for us to use AI hooks in our videos and I sent it to Andrew. And uh I know. >> It's so good. The the announcement from Higgsfield is introducing hooks by Marketing Studio, 25 plus hook variations for UGC videos, including cold opens, POV setups, pattern interrupts, and story hooks.

And that video that we just showed. And if you're if you're listening to this, uh if you're listening to this only, I'll link the tweet um in the show notes. So Mike, make sure that that link is in there, please. Thank you. Uh the um the the link up in the description so you can see with the video and the in fact, the link will be to the Higgsfield tweet about this. Uh because it's so it's so good as a hook. It's it's brilliant.

Um and I mean I mean, I'll just tell you where my mind goes immediately as I watch this, which is like gosh, should we go get that and start making these ads immediately, you know? Like like I think it does. It makes me immediately go like man, I want to do that. Let me tell you the second place my mind goes when I watch that. I have seen a thousand incredible uh individual AI instances tweeted as things that work awesome.

Uh and then you go and try and make it yourself and it actually is like dramatically harder and worse than that. Um especially with these visual type things, you know? Um so like almost like Twitter's always the highlight reel, you know? Especially when it's coming from the company itself. Like they chose that video to include and that tweet by the way has bonkers engagement, of course. Um so I haven't actually tried it yet, but this is like this is this thing I feel every time I see something like this.

The my very first reaction is the same I think as everybody's, which is like like that's incredible. That will definitely get attention and we should use it right away. The second thing I feel is like uh been down that road before and maybe I should just let everybody else tell me if it's like we'll just do it in a month when everybody tells me if they're using it or not, you know? Yeah, so maybe we can just I mean I got lots to say here.

We can just jump right into the topic, which is like how how to parse out the hype cycle versus like real tools that'll make meaningful impacts on your business. And so when I hear what you just said, you said like I see that and I want to go use it immediately, but we just ended a Slack conversation where you're like I need to avoid the hype cycle. And so my question is like how I I don't I can't make sense of those those two things in my head at the same time.

So how I don't know, let's talk through it. Help help me make sense of You can't make sense of my perspective, you're saying? Of of how those two things can both be there? Yeah, well, it's like we just finished a conversation about uh I'm you know, the more and more I think about it, the more I want to kind of avoid the hype cycle. And then there's this very hypy tweet that's like, okay, let's go do it and go do it now.

So how Yeah, help me help me think through that. Well, let me I mean I I have a lot to say about this, but what what do you What is your reaction to that video? What should we do as an organization as you see that? Um so, more broadly, I think when it comes to the hype cycle and all the constant AI hype and and it's not hype, like it's real development. Like every time you log on to Twitter, there's some new Claude announcement that's like really meaningful and impactful stuff.

And I, you know, I used to really honestly get overwhelmed by all the announcements and feel like FOMO and um you know, if I don't if I don't use this new feature immediately, I'm going to be left behind. Yep. Like that's a real thing. Like FOBO is the new term that's used, fear of becoming obsolete. If I don't use this tool, I don't master it, I'm going to become obsolete. You know, we as an agency will become obsolete.

And I I have felt that way in the past, kind of February, March time frame, but more recently, I've kind of realized that it's impossible to keep up with all of the It's just impossible to keep up with all of it, but >> Yeah. it it is possible and I think it's extremely important to keep a finger on the pulse of everything that's going on. So, now, when I see see a new tweet from Claude or whatever and I'm like I immediately resist the urge to think, oh, I'm missing out, I need to go do that.

And instead, I take a step back and I I kind of just put it in my back pocket and I think about, you know, how how can I you know, I'm constantly thinking through how can we leverage these tools that are coming out to move our business forward cuz we can't use all of them, so how do we pick the right ones? >> Yep. So, this one is a perfect example of something we talked about like 6 months ago and you know, neither of us have taken action on.

It's not anybody's fault, but you know, we have this idea we never went forward with it. Like we want to get do better hooks in our videos. So, how do we do it? And this is like this felt This one felt like kind of um the And I've been watching Higgs Field for months and months. And so, like I know I understand the features that they're rolling out. And so, this one in particular felt like an easy solve to a specific problem we have rather than So, maybe that's the answer to your question.

It's like how do you filter through? Well, we've got a very specific problem, which is we need to get better at hooks. Or we would like to get better at hooks. And there's a very clear easy solution um to it right here. And so, let's go try it out. Uh Okay, so here So, so I think two things about what you just said. Uh well, uh I think three things. The The first The first is I'm actually realizing that my internal reaction to that tweet, like what I feel about it, is the same thing I feel when I see somebody tweet um a Shopify screenshot uh of their success.

Uh which is Oh gosh, I got I better go do what they're doing right now. I better go do it right now. You know, because look, it's working for them. It works. But what I know about Shopify screenshots is that they are highlight reels. They don't tell the whole story of what's true. And uh and they're actually you know, there are actually multiple causes for almost any of them. The fascinating thing about this is I just tweeted like yesterday or the day before or something one of my most engaged tweets in a very long time, which was a Shopify screenshot.

And I I actually texted Taylor at one point. We were talking about it and I was like, "I'm sort of depressed by how well this tweet did because because it's a Shopify screenshot that is that is predicated entirely on media buying changes. And um and Now, the story was 100% true. So, it's I mean, that's why I tweeted it cuz I thought like what we did there was really helpful and it reinforced a lot of what I believe about media buying.

But Taylor said in an algorithmic environment where engagement is rewarded, everything eventually devolves into Shopify screenshots and porn. >> [laughter] >> And I was like, that's actually a like fantastic observation. Um and because like because the the sort of more nuanced thoughtful stuff like is is really hard to do. So, anyway, that's how I responded. >> to your tweet, there was more substance there than just a screenshot.

Like it was a great It was some great content. >> people people love the idea that like you're just a few tweaks away from massive success. And I think that's the thing that's the similarity to this. It's like what I feel about this is is oh my goodness, all of our ads will rip if we just have hooks like that, you know? Um it's it's it's a very similar feeling. But so so this takes me to the other two observations about your about the about what you just said.

One of them was you referenced being sort of thoughtful like asking the question how do we thoughtfully engage these things? How do we thoughtfully Yeah. >> like because because if you start from the premise that you can't do all of them, how do you thoughtfully in um create the ones that can actually make a big difference for us? I mean, even the point that you're making about about the fact that we talked about this 6 months ago is so perfect.

It it illustrates all of this stuff because what we were talking about at the time was using Sora 2 for crazy hooks. Yeah. And and part of that is because is I mean, frankly, because we were seeing screenshots uh or like tweets of like really cool Sora 2 stuff people people were doing. But when you actually go and do it, it actually takes way more work than it looks like. It it actually wasn't like one-shot super hooks, yeah, so far as I can tell.

And the the way that I know that is because it has been hard for us to try to figure out how to integrate it. First, we it is including like operationally. Like literally we have had to think about what Even I've talked about this a bunch. Like what operational space can we create in the organization for people to go and start using and doing this cuz the adoption actually isn't easy. Um and uh and so so there's an operational challenge to it first, but then uh uh but then uh on top of that what ends up being the case that when you actually go try to use these things it's harder than it looks and it's not successful looks and part of the way I know that is because I don't see that many people now saying oh man, we've been using Sora 2 and our business grew up, you know, 200% or whatever, you know?

Like in fact the the like the actual growth patterns and this is like all there's like what I'm saying is the business impact by these so far as I can tell is not actually massive. Um and and I think the way you know that is because you don't hear the stories of it being really massive. If anything, I actually hear the counter story which is like man, we're doing some of these cool things with AI, but um it's not super clear and I'll I'll actually reference some data later in this conversation to this end.

That is that is not super obviously making a big business impact. So you you reference you said in your comment which I agree with 100% that we that we're trying to figure out how to thoughtfully engage it. And this is like the shining this is the shining thing where I think you are like lapping the world on a lot of people on this stuff. Like when I watch what you've spent your time on um I I'm like it is it is all thoughtful and and get it is all thoughtful integration.

All of it. It's just like you're using it in these deep down process oriented ways that are foundational that we can build on top of them for a very long time as opposed to just like trying to go do a cool thing next. Yeah. And so the third thing that I'll say in response is um is you said like man, this would really help our hooks and it would be easy. And that's the that is the that is the instinct that I want to uh that is the thing I want to say.

That's that's what I mean by the hype cycle. That's what I that's that precisely is the thought that I want to be resistant to because in my experience so far almost everything that has come out you know, Claude tweeting about a new feature that you can go add to Claude code is not hype cycle, I don't think. Um I think that's like I think that's like flood has sort of earned earned the benefit of the doubt in this. Whereas like almost everything else, especially from a from other tools like where it just sounds like man, it'll work so easy 25 hooks like that.

In my experience so far, almost none of it has actually been that easy. It's actually been like rife with like re-prompting endlessly and taking my taking a bunch of time. And and and then like the final ones like so so you see that video that we showed and yes, if you can one shot that or whatever or even three shot it, you know, really really consistently then it is easy and it will work. But I so far haven't seen that and in fact the path to getting there it requires a bunch of work.

It requires somebody to go typically Yes. >> hone in on that tool a whole bunch to get it to be usable as at that level. So that's that those are all the competing things I feel about about this conversation and I'll leave that thought there. Well, I your last point I like when you say I know you weren't saying this but you know, you were saying it has you know, it has to be easy or it should be easy and it's not always as easy as you think it is.

And like what's going through my head is well, I don't care if it's easy or not. I care if it produces results and you know, for example, like all the software I've been building for three months has not been easy but it's completely changed our business. But when I see stuff like that coming out of Higgs Field, I mean it's pretty obvious. I know some people hate Higgs Field but it's pretty obvious that they're the leaders in AI image and video gen.

Like I don't care if it's easy or not. I want people one or multiple people on our team to be experts at Higgs Field and like on the bleeding edge of maybe [snorts] there's 80 different features in Higgs Field. There's probably a lot more than that. It's kind of overwhelming. And you know, maybe only two of them are useful for a business but but those two, a cat hook, you know, knocking over a bottle of, you know, those somebody's going to come across a one feature in Higgs fields that that's going to change our business.

And so, you know, I I don't know what my point is other than like, yes, there's a hype cycle and yes, there's all these features constantly being released, but this is like the new state of business and the new state of play that we're living in. And if if we as an agency are not constantly plugged in and and learning and testing and uh exploring the new tools available to us, not probably 90% of the time we're going to find that uh that really wasn't worth the squeeze, but the 10% that we do find, I think is going to be really game-changing for us.

And then the the the other piece on top of that that I think a lot about is if we're not doing it, some other 18-year-old kid in his basement is going to do it and figure it out. And, you know, maybe this is me having a little too much FOMO. Yeah, that's that's exactly what it sounds like. What what you're expressing I was just going to say this, like it sounds all all I hear in that is FOMO. It's just like well, it yeah. >> It's an acknowledgement that like um I mean, we we obviously have a very clear belief that you know, a kid in a basement with Higgs field isn't going to outperform us as an agency.

Like there's a lot more to the work we do. But >> Yes. But uh he or she could offer a creative business that might rival ours and then kind of slowly chip away, you know, at at um at what we do. So, I don't know. It's not like I don't want to constantly be uh making my decisions on what we do as a business based on what I think everybody else could or might be doing. I don't want to do that, but I also want to pay attention to um like the the analogy I use is, you know, what if what if like 20 years ago or whenever Adobe Premiere came out, you know, and we were running our agency and we decided, you know, we're not really going to do that Adobe Premiere thing.

We still like cutting film clips by hand and assembling them together and like that just wouldn't like I think about it's kind of a silly analogy, but it's the right one because the the the state of play is changing right before our eyes and if we don't keep up with it, the guy who is using Adobe Premiere over our video clips is eventually going to replace us. And so, that's a dramatic way to put it, but I do think the the seismic shift in the industry in terms of the tools that are being used and everything is being upended.

I think it's very I think it's I think it's even more um disruptive than the analogy I just gave. Maybe. Maybe I'm overstating it. Somewhere in this episode, I don't know if it's before or after this ad, Patrick mentioned landing pages as part of what AJF Growth is continuing to do. And we are as part of that, partnering with our friends at ECom Experts to get those landing pages integrated into websites when we need to.

If you've done any even simple work like that, uh like like landing page development work, you know that there are there are even if the page is relatively simple, there are often, especially if you want to mess with the offer at all, all of these little things that come up that can make the whole process break for customers [music] and all the rest. And so, you need a web developer at some point to take your idea from the visual design to an actual smooth clean integration with your website where uh where [music] add to cart buttons work and offers work and checkout works and all those kinds of things.

And ECom Experts has become a great partner for us in accomplishing that because they are a great performance uh uh web dev agency. And what I mean by that is they can handle all the sort of technical integration things that are just part of web development on a Shopify store. At some point you just need somebody who can code something. And in my experience it's been very hard to find reliable [music] people who get work done at a reasonable cost on time.

And so I'm constantly referring to Ecom Experts. In fact, the way that we got started with our connection to Ecom Experts was was that we were working on a client together and they were surprised to find a growth agency that seemed to be doing a really good job as you know can be really hard to find that. And we were surprised to find a web dev agency that was doing a really great job. Everybody was very happy about it.

We went, "Wait a minute, we should work together more." Because in my experience so many dev experiences are just so bad. They're so frustrating. They get delayed. They get too expensive. They get bloated. All of those things. Or they they web dev projects get done with just no mind for performance and for conversion and that kind of thing. And that's one of the great things about Ecom Experts is these these are like ex-Shopify people who are not just dev people, they're people who were who were building [music] stores to sell products.

They understand that you need to do that. So they can take over your full CRO experience. They've done it for huge brands. They've worked for a long time as sort of a right hand to Cody Plafker and Jones Road Beauty. They just understand this stuff deeply. So between performance, design, CRO, ongoing testing, and then the sort of technical integration elements of web dev work, Ecom Experts has just been a great partner for us.

We expect to keep developing that partnership over time. [music] You should explore it too if you need web dev help. Whether it's ongoing support, CRO type stuff, landing page work, or even a full site migration, you should get on a call with them. Really big fan of there's ecomexperts.io, ecomexperts.io. The link's in the show notes. Go check them out. I think you might be. I actually I think eventually is a really important word there, which is and is and is critical because actually this is something I've been that has like lessened my phobo is watching how much we are smashing for clients right now and just seeing how many clients are still coming to us with problems in their businesses related to all kinds of things that we know how to solve.

We We can't fulfill the demand right now, right? Like we we and I don't even mean that as a brag. I mean I mean when I try to put my attention on the actual business outcomes that we are producing for both ourselves and our clients. Uh it's not that every business we're running is like completely blowing up because that's and this is I think an important thing. That's actually just not reality. That's not how businesses work.

No business There is no magic secret to making it so that your business will automatically just like completely triple overnight. There is no hook generator, there's no AI writer, there's no You know what I mean? Like all of those things it's just like I mean it's possible that at some point, you know, some business will do that, but there is no thing that does that. What what What actually has to happen for businesses is they have to have found do foundational things right consistently, you know?

Um and and so because of that I think I think part of the frame through which I'm I view a lot of this is like well if I if I want to go back to your comment, which is like how do we thoughtfully integrate these things? The the answer to me is that you relate it back to the core drivers of success for the client and ultimately for our business relative to the constraint of the goals we set, right? So so we have goals around how many clients we expect to add uh this year and next year etc. and at what kind of margin we can do it.

And right now, we're actually ahead of pace on those goals, right? So um so we're in a spot where where we think we can probably beat our pace for this year. Um uh but with with client additions even even without having to like massively speed up. Okay, so there's a constraint on how fast we have to adopt these things. Like right now, is any client going to fire us because we're not fast enough on AI adoption? And there's a couple ways that would happen.

One of them would be explicitly, like they would say we want more AI adoption so they fire us. Another one would be implicitly, which is that our performance isn't good enough relative to what's available for them at a relative price, and so they go somewhere else, and they would never call it AI adoption. They would just call it performance, but that would be the driver of it in a way that's sort of nobody could see.

And I don't I just don't see it right now. Like I when I look at what we do, I just don't see that happening because um because because especially I think because of the combination of of well-honed strategies at this point that um that I believe in and that, you know, realistically flow down from my mind, that have been operationalized at a level that I was absolutely unable to do before um because of the work that you've done building out AJFOS.

And so um so then the question is like when I think about something like Higgsfield and this this video, my response my response to it is, "Okay, that's pretty cool. If that could work really, really well, um let let's definitely not plug our ears and be like, you know, Amish about AI, you know. Um but I I I have no interest in that perspective. AI has transformed our business and I and I think that I think that we are on the bleeding edge of AI um adoption.

Uh and and so I want to do that, but I don't but what I don't feel is a need to be first. I think that's part of it. I don't feel like the Higgsfield thing is I Once I sort of reevaluated, I'm like, "Okay, let's find a way to integrate that. It's going to require some process-oriented things. Is there Is there some way on our team? Can we even build a process for like um like I've said to you, like a skunk works process, you know, like is there somebody whose job it can be to be the person who goes out and tries new things and figures that out or something like that?

But then in some way it's always related back to the core mission and to the core and to the core business goals and some of that stuff. Um I think that's what thoughtful integration looks like is that it's in some way related to our clients' goals and to our goals, and if we are hitting those goals and timelines with some some of the possible risk of those things always in mind, um we say, "Okay, we're going to keep on we're going to keep on just doing this a little by little because then that eventually won't ever actually come." Yeah, I mean for this topic in particular, I'd rather just plug our best artists into Higgs Field and just say, "Go do great things, you know, and tell us what happens and, you know, maybe maybe we'll identify like really core clear things that we are think think are foundational." So, back to your comment earlier, like what are the foundational things?

Well, in a video, a hook is pretty foundational. So, one of our artists may come across, "Hey, these hooks are consistently winning better than others and here's how we make them" and then roll it out to the whole organization, maybe. Yeah. So, I'm I'm hesitant to make any like processes or like, you know, SOPs around how to use Higgs Field hooks, right? Cuz it just came out and we don't, you know, I'd rather just part I'd rather give great tools to smart people and and say, "Go figure out the best way to use them." Yeah.

Um but but we but then we do have to create space relative to their other requirements to get there because this is part of the point. This is part of the point, right? Is that like currently doing the things that we do, we're actually delivering really good work really consistently for our clients. And so so there's a real cost to having that person go and add Higgs Field, which is they're then going away from the thing that we know works, right?

They're they they have to actually do less of that thing that we know works. So, there has we have to believe this thing can can be better and I believe it works because actually because the message is the most important part of the ad and that there is and that like hooky replacements for the message only gets you so far. And so and so like that's that's another lens. It's like a first principles lens through which I view something like this.

And and I'm going to hold that as hard as I can that like just resist the urge to go find something else. Now, I could be talked off that if there was data showing me that, but right now I don't see it. So, um so yeah, the the uh but yeah, that's that's part of the thing I feel about this. The problem with give them the tool and have them go do it is like, okay, at the cost of what else though, right? Because especially if it's not going to be as easy as it looks.

And and by the way, one other thing on this, especially if there's a new announcement like this from somebody every day. Yeah. And that's a real thing I feel, right? Is that like And that's why I like the idea of like a skunk works kind of person is like, okay, who can actually be the like evaluator? Right, skunk works skunk works You You worked for skunk works. Do you want to explain what skunk works is to people so they know what the reference is?

Peo- peo- people on your team who are dedicated to like advanced research and development, you know. So, um Out- outside of the core processes. Like >> Yeah. That's That's why I I mentioned a couple weeks ago we were talking, I wanted to talk about putting making somebody like a I called it a creative director, but you know, that that would be the function of that person. Yeah, that's a good idea. Um Well, you told me it was a bad idea.

Well, you didn't say that as part of their function. I think that's a smart idea for what their function is. I'm pretty sure I did, but Oh, well, maybe I missed it. Uh maybe I didn't. Um okay, let me I don't know how real we're getting on this podcast, but let me ask you what you want. Let well, it's this is business real. So, Oh, okay. We can always sleep it. No, let me let me frame Let me frame a question to you. Okay.

Or or let me try to explain one of my phobo fears in like real terms. So, we have a client who's a media buying only client. We only do media buying for them, and we don't do any creative forecasting. And as you saw a week ago, uh they're cranking out chat GPT-2 uh or sort I don't know what they're calling the model, but the recent version of chat GPT image generation is basically I mean, it's it's on it's it's phenomenal.

You can do whatever you want with it, and it's it's amazing. So, we've got a client who figured that out, and he's cranking out he's got the, you know, infinite ads machine that he's already made. And is it great? You know, how good is it? I don't know. Let's not talk about how good the ads are performing. Let's just talk about visually, stylistically, message-wise, you know, you can have these ads do whatever you want in one shop.

Yeah. Why why on earth would that client upgrade services for us to go make to to to hire us to do their creative? Like if they're if they're And I'm thinking if I'm the client, like I've got a little button here I push that makes really good ads and and and they're working great. In fact, they work amazingly. I see no difference between my ads and these other ads you're going to charge me, you know, 10x for. I don't know.

I don't know what we would charge more, but you you get the point. Like why would they So so my question isn't why you know, you you framed well nobody's going to fire us because we're not adopting AI. That's not necessarily my fear. I think that that would be silly if somebody said, "We you know, Andrew, we just we don't like you guys cuz you're not using enough AI." Nobody's ever going to say that. What they're going to say is Well, I I don't want to pay you for this cuz I can just do it myself and I'm going to click a couple of buttons.

So like what >> Right. How do we stay ahead of that? >> do think that's a that's a concern. Yeah. How do we How do How do we respond to that as an agency? That's kind of Well, they I mean you you said what the answer is in your question, which was if if our ads perform the same as those, then uh then they should not hire us for that service. Like I think I think you're just absolutely right about that, you know. Um I just don't believe that.

I just don't think it's real, you know? Um like Uh yeah yeah, I mean Yeah, I like I I think it's possible that it's going to be real soon and if it is then it's going to be a a real problem, you know, for us, but um but I think I think uh and I or or we'll have to change the service in a significant way. You know, I I think this is unquestionably one of the biggest risks in our business is that is that stuff that we can charge for right now the cost goes down to zero.

Exactly. I think uh I think that's definitely a possibility, you know? Um and so the but I I'm not actually sure that staying up on staying up on AI tools helps solve that problem, right? Because if that's real, then they'll just do it themselves either way. Well, there's a price there's a price they would pay. Like if they're like you're charging me 200 bucks for this, I'll pay it. Like I don't know. I'm making up dollar amounts.

Like yeah, for for me to not have to do this and you know, you're an expert and like there is a price people are willing to pay. I think what I'm trying to get at is like you know, and we talked about this a lot. Like you know, the model gets squeezed and squeezed until there's you know >> Yes. But you you actually know what my answer to this question is. Mhm. And you hate it. Which is? Start brands. >> [laughter] >> The answer is if it be if yeah, I mean this is this is legitimately part of the reason part of the the thinking for me of that of that thesis, which is like >> Yeah. if suddenly all of the value like think about that.

What you're saying is that the margin value evaporates from the service business. Well, who does it accrue to? It just it just accrued to the client that you just described if that is the case. And so go be that person especially if we're much better suited to run an e-commerce business for all of the other things that AI cannot do a one button thing for, you know? Maybe or maybe maybe you change your service, you know, maybe Not for $200 a month.

Yeah, I mean obviously. Yeah. [laughter] Yeah. I have no interest in that business, but like um and Okay, here's here's something they can't do on their own right now. Uh make landing pages the way we make them. Yes. >> Uh do you know, like influencer marketing, creator management. Like there's so many services that I I think >> need creators or influencers if the if the ads perform the same. And and they don't need the landing pages either if the ads perform the same.

You know what I mean? Like Like in everything you just described, it's like what's the point of paying for those things? If I can just you know, if I can just make endless ads that work. Well, yeah, I mean you need landers and I mean you don't need them, but uh Yeah. in in our I don't know, in my experience you know you need a good lander. I mean we've seen a client Won't those Won't those In the world where a client fire agencies for performance creative Right?

Like I'm just trying to I'm trying to I'm agreeing with your with your principal thought here and trying to chase it down further, which is to say in your world that you're describing where performance creative is that easy right? In that world, there's I have no interest in the lander business because the lander business is going to be one shotted 3 weeks later. Yeah, I think what I'm I think the point I'm trying to get out is our business requires This is where we really started this conversation, requires us to to constantly have our finger on the pulse of the kind of latest and greatest technology and constantly be adapting and constantly reassessing, is this a business that we can be in long-term?

Like we need to be having conversations of is static creative a a um Yeah. strong strong margin business that we can be in. Like right now, absolutely. For this for the near future, definitely. >> Yeah. I don't know, next year, I'm not sure. Um and I think we can't have those conversations unless our finger is on is constantly on the pulse of what's changing and what's happening in the industry. All of the creative that Patrick and I and our team is making is powered by my friends behind the scenes studio.

Behind the scenes studio is a creative agency that is housed in the Philippines that works directly with me and my team and is is really downstream from a lot of my best ideas >> [music] >> and is taking those things and applying those ideas to ad accounts that we are not running. Uh they do of course more than just what we do. They do all kinds of design and video ad work for e-commerce brands. But if you like the ways that I talk and think about creative and you need design and video editing help for the ads that you are making for your Meta ad account, you should consider working with Behind the Scenes Studio.

AC Gardner and her team are really awesome to is fully built in the Philippines. They are approachable at their price point. If you need a bunch of creative but you're not quite ready for a full performance agency and you you know you want to handle the media buying in house. You just need the creative extension of your team and you want to think the way that I think about ads, you should get on a call with Behind the Scenes Studio.

That's really what they are and what they do and I think it's a great option for a lot of brands that frankly we can't work with. We refer them all the time and and people are really happy with the work that they see from AC and her team there. [music] Um so if that's what you're doing, if you're trying to increase your creative output for your brand, you're trying to integrate AI into your ads, you should do it with AC and her team uh at Behind the Scenes Studio.

Go to btsstudio.co. to get on a call with Behind the Scenes Studio. See if their creative solution is right for your brand. That is all they are. They are built entirely around making Meta ads accounts grow with performance creative. Go check it out for yourself today. Okay, let's let's enter into the evidence base for for this conversation, another another set of voices, okay? Um and I let's let's evaluate these together Great. >> that I think is really fascinating, okay?

I'm going to enter into the conversation Brent Beshore and Andrew Youderian. And actually in both cases it's really not them individually. It's Brent Beshore's 16 portfolio companies and Andrew Youderian's 300-plus e-commerce brands. Both from 2025, which I think is a real is a real question here, which is really interesting. Okay. Andrew Youderian e-commerce ancient, possibly. Yeah, possibly possibly That's possibly the answer here.

Um Andrew Youderian trends report and I'm going to link my conversation with him, too. So, I did an interview with him about his e-commerce fuel um, trends report. The My full conversation with Andrew is in is in the show notes. Uh, so you can go get back to that cuz it's it's one of the coolest things to come out of e-commerce every single year. Udarian runs the e-commerce fuel. Great dude, you know him, hopefully. Um, and and he does this survey of 300 e-commerce brands, okay?

Um, I did a full conversation with him breaking down the survey. Uh, go get the link for that in the show notes. And by the way, while you're at it, subscribe wherever you're watching or listening. Leave a comment on this conversation if you are wrestling with these things. I would love to hear how you are thinking about this in your organization. Um, your adoption, what tools you're using, anything. Whether it's as practical as the tools you're using or as um, or as big as sort of your broader reflections on this stuff, the way that Patrick and I are working through right now.

Please leave a comment on the video, and I would love to read and interact with you there. So, and subscribe, like I said, wherever you're watching or listening. We got more conversations like this coming all the time, okay? So, if you like this conversation, if you're this far into it, you're going to like my other stuff. Go subscribe. All right. Andrew Udarian. Um, Udarian says in his report, um, uh, here's the headline from his blog post about it.

AI is amazing, but it isn't generating ROI. From 2025, revenue growth for brands that were AI adopters versus non-AI adopters, self-described, okay? Um, in 2025, were virtually ident- identical. Actually, the AI adopters were a little lower. Revenue growth 27% for um, for adopters, 28% for non-adopters. Net income growth, like actual margin profit growth, at non-adopters of AI grew 55%. AI adopters 33%. So, that actually the non-adopters were out-margin the AI people by a bunch as a percentage.

Um, now, the net profit margin for the AI adopters was higher, but the growth percentage was faster for the non-adopters, okay? Um, revenue per- revenue per employee for non-adopters was higher than it was for AI adopters. So, the AI adopters were actually having less revenue per employee by about $100,000, uh 930 to 830, roughly. Um and the non-adopters had one more employee in the business on average as well. Um and so, uh this controlled to an uh the control the data here was controlled to include profitable businesses with 1 to 25 million dollars in revenue.

Um uh less than 100 100 employees. So, it was like comparing like-to-like in those businesses. Um so, um so, really, really fascinating. Basically, that's uh that's basically 180 companies who that's Oh, n equals 192 there. So, 192 of that of of that group. Basically, um basically, it was not making a definitive business impact. Okay, then, Brent Beshore in his annual letter in 2025, 16 portfolio companies, private equity.

He says he added a PhD technologist at the top of their companies to help uh AI adoption. So, the person was really smart, really, really talented. And across the board, they saw no impact on their business from AI adoption in 2025. Uh and then, said when he talks to other friends of his in the space, um uh that they they said uh they said that it was basically all window dressing and it was not it was not uh it was not at all valuable to them.

Uh in his in his annual letter, I'll link the annual letter as well, really, really good annual letter uh even if you end up disagreeing with his perspective on this. Um I'll I'll link with an anchor link straight to the uh straight straight to the part of the the the letter that talks about this stuff. Um but, uh but, yeah, while recognizing that um every week his input here is changing, right? Like, all the time. Um uh or rather, the the the the technological world is changing all the time.

While recognizing that, they were finding that um the quote that let's see. Here's Here's the full quote. I recently chatted with the founder of a large private equity firm that has publicly touted their AI program in hopes that they figured something out. They haven't. When I asked him about the value of their AI efforts, he said, "Quote, absolutely nothing. It's window dressing." Okay, so So, that feels insane to me because I'll just say that's simply not true in our business.

We could not be where we are without AI adoption. So, the idea that it's absolutely no value is simply not true at AJF Growth. It is not true. Like we are growing our business at the level of profit and revenue both because of that. Okay, so just definitively not the case. Um Uh but the thing is I think there is a chorus of voices who are singing this song, which is there's a lot of promise, but it's not super clear that it's made a super big impact.

Now, maybe the answer to that is that was last year and this is 6 months later and uh and that's the entire answer or 5 months later, but um I don't know. What do What do you think about those people saying these things in the midst of this? Uh Well, tell me, what were you doing a year ago with AI? What were we as an organization doing 1 year ago with AI? Yeah, yeah. I mean, not much. Using We We had plenty of face. Chat with some prompts, you know, it's like I I don't mean to sound dismissive, but like 20 like it's that's just ancient news. 2025 is ancient news.

Like the the step function change in AI was at the end of December 2025. So, it'd be like saying the printing press was invented in December of last year, but you know, no printing companies have seen any benefit from the printing press last year. It's like, well, that's kind of silly. Um I mean, and even think about our organization. Like when you think about Oh, yeah. what AI adoption made any difference for us in 25?

And like when you think about 26, like it's completely changed our entire business. Um >> Yeah. So, I mean, I I think those guys are smart and I I like both of them and uh but I just think the reports are like not really helpful for for like May of 2026. But doesn't it at least push back against the hype cycle a little bit? Because the hype cycle was pretty serious last year. Yeah. >> Like it at least at least at the very least to me what it says is there is at times a real difference between what people say is happening in terms of the impact of AI and what is actually happening. >> That's absolutely true.

And I think that that is like the big contribution of it to me. And what it rein- I'll tell you how I interpret it. How I interpret it is resist the hype cycle. Instead, find ways to integrate it not as a be-all end-all solution to everything, but as a way to enhance the things that we believe work consistently and steadily over time. And don't worry about FOMO, don't worry about the hype cycle, you know? That's like that's the big contribution of it to me is like Yeah.

I mean, it's also probably worth saying that like some of those investments just take time to pay off. And so it's possible that that will change a lot in 2026, you know? Yeah. I think that's also really reasonable. But um But it it it is a remarkably strong statement from some of these people. That's the one of the things that strikes me about These are smart high-integrity people saying something that is that is that is they're saying it they're saying it really strongly. >> Yeah.

And I think it's a I think it's an interesting thing. Well, I mean, for for Brent, I don't know what his portfolio has consisted of, but it's not it's not like tech-native companies like us that are like depend on technology for every single waking moment of the work that we do. You know, I know he's I don't know. I don't know I think he's got more kind of blue-collar companies that like it makes sense that a Yes. It just makes sense. >> Yeah, like a manufacturing company's going to probably go uh you know, we make widgets, doesn't make a big deal for us.

Like, you know, um it's just it's it it's hard for me to take seriously anybody that says AI hasn't made a difference in my company. Um because like it's just so obviously true to me that it's completely changed everything. Like it's just so true to me. Um because I've seen it in my work, in our work, in our business. Like that Here's Here's a data point. Um you know, you just mentioned we in on a call that we just got off of, highest revenue, highest profit month ever just booked for AJF Growth.

Like Yes. I It's hard for me to say what I'm not saying cuz I don't think it's true, like 100% true, but like you know, everybody has been a part of that, but I don't know that that would have been possible if the tech piece that I've done has not had Like there's no way we could have onboarded and launched and serviced all these clients in the way we've serviced them if it hadn't been for all of the tech that we've deployed over the last 3 months.

Well, this is where This is This is where I'll say like and I have said this before, it will happily say it again. I think it's a there's a pretty solid chance that AJF Growth would be dead if you didn't come. Yeah. Yeah, so So like fear of uh of What is it? Be Fobo? Uh becoming obsolete? Yeah, that I think was like I should have feared that more last year. So [laughter] Yeah. Yeah, yeah. Like I think that's So my counterpoint to to your reports as well, all I know is my own experience, which is it's completely changed our business, led to, you know, our best success that we've seen so far, and from what I've seen, I've only seen the last Well, I've seen a lot, but like from what I've seen over the last 5 months of progress, like and what I know about technology curves is it's just it's going to it's exponential from here.

You know, any complaints that we have about AI are just going to be fixed. They They just are. And um I don't want to put all of Like I want I don't want to make it sound like I am a believer that AI is going to save all of us and fix our business and or whatever, like solve all of our problems, cuz it's not. Like I had this conversation with one of our employees yesterday, and you you know who that this one is, and he {slash} she, whoever it was, said, "You know, I'm just afraid you're not going to need me anymore pretty soon." And I said, "Look, if you promise to never quit, I'll promise I'm never going to fire you.

I need you. Like I need humans in my business. I'm not firing you because I think a robot's going to replace you. A robot is not going to replace you. I'm confident of it." And um but we have to pay attention The other thing I said was, "It is obvious, it's in the front pages of the Wall Street Journal every morning, that layoffs, jobs are getting replaced, you know, the the world is changing, and you know, we got to figure out how to adapt with it." >> Well, so this is funny though, because on the one hand, the three big layoff companies uh now who are who have cited this, Derek Thompson just tweeted about this.

The three big ones are Block, um Coinbase just did it, right? Coinbase just said that they are doing that as well. Um and there was one more that cited um >> Was it Salesforce or I don't know? Yeah, yeah. And basically And basically what he called it was AI-washing the the layoff announcement. >> Oh, yeah. It was I think it's a good phrase, because the point is like people looked at Block and they were like like if there's one thing you know about Jack Dorsey from Twitter, and if there's one thing you know him from looking at them, is that they were like dramatically overstaffed Yeah. >> post post-COVID in particular.

And and that like Yeah. And that like basically everybody has always thought that about his companies. >> Yeah. And and now he's just doing And now he has a he has a way to say it differently, you know. Um you know, Coinbase, it would not be at all surprising that any post in the post-Bitcoin hype bubble or whatever, you know, whatever you want to call that moment with Bitcoin, like people are not talking about those things in the same way, even if Bitcoin is a technology that will last for a lot longer.

Post-hype, at least, you know, it wouldn't be surprising at all that Coinbase needed to get a little smaller, you know? So, um so there's that. And on the other hand, Jensen Wang Huang, I don't know how you say his last name, Nvidia, uh Jensen said, "I don't think AI is going to replace any jobs at our company." Like he [laughter] he's saying the exact opposite. He's just like He's like he's saying he's saying I think it will only help.

It's going to lead to more employment, not less. And it look, you can you can say, yeah, like maybe, but I think I think you have to take both perspectives with a grain of salt, right? Like you have to say you have to say, on the one hand, this the everything is going to break because of it, you should be suspicious of. On the other hand, you should be suspicious of it's not going to replace any any jobs in our company.

You know, like I think probably I think they're probably both not totally true. Um yeah, that's the way I think about it. The way I think about it is we're not I mean, for us, we're not laying anybody off, we're not firing anybody, but I can say for certain we are going to expand our services and the amount of work we do, and we're going to do it with less people. Yeah. And like as we grow, we won't need to hire people that a year ago we would have had to hire.

Like right now, we're launching we're basically have a landing page service that we just launched that a year ago we would have had to hire at least one, if not two people to run. And we don't have that. I've got an agent that runs it. Um and that's going to continue to be the case as we grow. Uh and you know, frankly, I think we again, I don't know how real we're getting, but I think we should have real honest candid conversations with with like, let's just say, um with team members who whose jobs may be affected by AI in the future and go, "Hey, we're not firing you.

We don't want you to go anywhere, but like this is what we see coming down the pipe. So, how can we help you grow as an employee into a different kind of role?" Yeah. Or you know, maybe maybe what happens is as we grow and have more clients, maybe we just have less people per pod, you know, and yeah. I don't think we'll ever get to the you know, the CTC uh model, which is just one person per pod. You know, I I just don't see us getting there.

Um But I But I do see us doing more with less. That's probably the best way to say it. So, I'm I'm mixed on this because nothing absolutely nothing in all of this has made me think for 1 second that the core thesis that we have about client services, which is that great work requires requires deep work basically, right? Like this idea that we we go deeper with fewer as like a distinctive of AJF growth, right? We have three full clients per cert per brand unit.

I kicked that around on a podcast at one point that that each each of our growth strategists only has three full clients that they that they're like really truly deep with, right? Uh they might have a little media buying client on the side or something like that, but but uh but weekly meetings, full creative services, full cohort, fat forecasting, all that kind of stuff. Right now, our growth strategists have three clients, and there are nine people on those teams.

So, there's there's a huge huge amount of human effort going towards a very few clients. In fact, while we are adopting all of this AI, we are doing that in a way that we think is distinctly more human-centered than any other agency in the world. Like I don't know another performance agency that's less than five um clients per pod, right? We're at We're at We're at three, and we have nine people servicing those those uh those clients.

So, we we actually are going in a like just already, we are distinctly in a different world than what so much of um the our space is. And I so far see nothing about the businesses that we work with, especially as they grow, that makes me think that we do not still need a huge amount of human power and a huge amount of depth to make those businesses great. Um and uh and this is what This is again where I like the principle is a constraint to me in a way that like until I see it to Yeah.

And And And here again, I'll enter into the evidence, your own testimony about yourself, which is like this was all incredible and also it broke you. You know, like like trying to do that much work that fast with AI like cuz the thing that hasn't changed is the human and the human capacity. And so like and so like you know this is I think where this category between or this trade-off and and I'm I'm not trying to make this sound like an easy an easy thing for me.

I'm trying to say I'm really wrestling with this because because I see everything everybody is saying about all the change and the disruption. I see all of it. Of course I do, right? And I'm just trying to figure out like what is actually real in our business right now versus what is not. Like how do we serve our client today? How do our employees be great today? And and and also how do I look down the pipe a little bit and or look into the future a little and say how do we not die?

Like both of those things are true. So So I uh So but right now if I'm as honest as possible with the evidence, our current model of deeper with fewer a bunch of people like that is producing the best work we've ever done at AJF Growth for our clients and the most revenue and profitable month aided by AI gigantically, right? Like built on a foundation of AI. And I think so right now it's like well I guess that's my thesis right now.

So that's what you can tell, you know. Yeah. I think it goes back to where we started which is and this is what has been our kind of guiding principle and it's served us well which is uh solve a problem. Like identify the problem first and then try to solve it. That's right. Where I think it goes wrong is and we've we've been guilty I've seen this in in our Slack channels. Like oh hey look at this new tool, test it out, you know, try it out.

Look at this new announcement announcement and like even I've been guilty of just flinging out announcements and Yep. you know, like there like Yeah, there's a difference between just going hey this cool thing just released, let's use it and going hey we've got a problem that we need to solve and we think this can solve it. So that's why I sent you that. Going back to the top of the the call, the Higgs field thing it seems like a really elegant solution to a clear I won't say it's a problem we have but a clear desire and a need we have.

So, like that is like a perfect example of um something that we could go we could go do. So, um yeah. If it actually works the way that it says it's going to, right? Which again, I this is like where there's still this bit of skepticism for me, where I I feel burned by the AI companies and by the marketing at this point. Like I mean, just think back to icons announcement, right? Yeah. >> That's the most extreme version of it.

Like, we're just going to replace everybody on your team for forever. And it's like, oh, actually they're bankrupt and dead. You know, like Yeah. you know, so it's just like It doesn't mean it won't happen. It's maybe we're just too early, right? But like, yeah. Yeah. Yeah, I think All right, we're out of time. Okay. We're out of time. We have to save more of it for later. Um I like I like that last note to end on, which is which is like actually trying to be thoughtful about what the work is and what the problem to solve is and being that being the lane for how to do this, um I think is is uh is a really good thing.

And I And I should also tag this by saying one more thing, which is that we don't have a team of people who are resistance to AI. Uh like our designers and and artists and stuff are like are trying thoughtfully to to figure out how to do this. They're doing a really good job with it and it's really fun to see. Like, we can always do more, but they feel it like we do, but um but just to shout them out. They they do great work and and, you know, I'm looking at ads in our ad account that are like AI designed backgrounds with very minimal assets that are like the top spending ads in the account, you know, really cool stuff.

And um and it's it's fun to see. So. Yeah. All right, man. Thanks. Thank you. >> [music] >> All right, you got to tell me what you think about this episode. I think a lot of brands and a lot of companies in general are wrestling with the stuff that Patrick and I wrestled with here. I would love to hear your thoughts. Please, please leave them in a comment. Like I said, I read every one of those comments. I interact with as many of them as I can.

Leave a comment. Tell me what you think. What did we get right? What did we get wrong? Where am I an idiot? Listen, it's very possible I'm just an idiot on this. Please tell me. And uh of course, email me podcast@jeff.com. If that's the easiest place to respond, that's that's fine, too. Uh I'd love to I'd love to get those emails and hear from you there. I'd I'd love any questions, any thoughts. If you have an episode idea, if you want me to talk about something, hit me with it in email.

So, big thanks to Patrick for taking the time to do this. Of course, if you want to work with us and the AJF Growth team, we don't really have any space right now, but we probably will soon at some point here or hope to have some soon. Go to ajfgrowth.com, fill out the intake form, tell me a little bit about your company, and I'd love to hear from you. So, ajfgrowth.com [music] is the place to go to get all of those things.

Subscribe wherever you're watching or listening. All kinds of great episodes coming soon. I recorded my episode with Phil Keel on CPMR. If you're interested in the media buying weeds, talk about that. Got Ezra Firestone coming very soon as well. It's going to be great. Thanks so much. Talk to you soon.