AI Is Devaluing Your Agency | Episode 210

Jake: It seems like agencies lately have been devalued in the eyes of clients. Now everyday people and business owners have access to literally anything that they could ever want.

Co-host: That's crazy. Like, that's thousands of dollars that I just saved because it became accessible, that I wouldn't have to pay somebody a, uh, professional service fee. Nobody's paying for entry level content writing. That's just dead. It's like dead as a market.

Jake: Now that people have access to these tools, there's no longer a need for simple tasks. We don't need a dev agency to set up a simple API anymore. We don't need a designer to make simple Facebook ads. We can just do that. And if we don't need that, then customers don't need that either. It's all about what's happening in the consumer's mind, not about what's actually real. He just had an upset client because he sent over a design that was used with his actual design software.

Co-host: Wait, what?

Jake: The customer was upset because they did not use a, uh, design generated by ChatGPT. They used their own design from their own design software from their own brains. They just said ChatGPT knows everything and therefore the designs are better. Be honest though, I bet you didn't go to your lawyer first, that you probably went to Gemini or ChatGPT or.

Co-host: The amount of other stuff that I've done without a lawyer at this point would scare any lawyer for sure. And it has worked out.

Jake: You also took a really big company to court and yeah, that happened. And you consulted with Gemini the entire time.

Co-host: I didn't know what I was doing. I didn't. Yeah, it just walked me through the whole thing. It's crazy. I think so. I want to know. I want you to talk about it more like I want your thoughts talked through it. Did that make sense?

Jake: Yeah. I'll do it briefly. So, uh, I finally published the Big Schema SEO markup report. That's not English. I published the big SEO schema markup research, uh, study that we, uh, had been doing since almost January. And I published it on our website today at 11:00am Central Time and it is 7,135 words. WordPress estimates it at a 38 minute read. Um, it's got everything in their methodology, like, um, more screenshots of the data and everything that you could possibly want to do with. If this was faked, it would be more effort to fake this than for us to just pull the data, um, and paste the screenshots that we put. Um, I think if I pick this

Co-host: up, you're going to get an honorary degree. Somebody's going to give you one. That's a dissertation right there.

Jake: So I had peers review it, and I called it. And they're experts in the industry. They're, um. Let me list off some of them. Um, there was, like, Jarno Van Driel, who, like, invented schema. Not literally, but he came onto the schema.org team in 2011, right when the partnership with Google started. Um, and, uh, he consults with, like, John Mueller on schema and everything. We had Yoast, the principal, uh, SEOs at Yoast, uh, review the schema markup on the site or on the sites during and after, um, the control and test periods to. To validate that. Yeah, it was placed. Um, we had David Hunter, CEO of Local Falcon, Jonathan Berthold and Tom Capper from Moz, review everything, the methodology. And then, uh, Darren Shaw and Joyce Huckins. Um. Oh, I put Joyce Hawkins in here, isn't it? It's Joy.

Co-host: This is Joy, right?

Jake: Yikes. Hold on. I gotta edit this real quick.

Co-host: Moving too fast and breaking things.

Jake: Yeah, she hasn't said anything yet, which means that she hasn't even read it. Um, which I don't blame her.

Co-host: Um, it's gonna. You're gonna have to have some real nerds that are gonna read a full 40 minutes of schema research analysis.

Jake: It's already updated. Um, but either way, theory, the methodology just didn't really give a whole lot of input. But, uh, I published it, and now I'm looking at trying to get it published in Search Engine Journal, which I guess is harder than Search Engine Land. Um, and, uh, I'm publishing this version of this in Moz also. So this has been, like, a really long, long and hard effort. Um, like, we've been doing this again since January, just, like, recording data. And then finally, um, In April, like, April 3rd, I think all the data was done, like, done collecting. And then we just had to, like, analyze it. Here we are on June 3rd. Published. Finally. It took me two months to write it.

Co-host: What'd you learn? Um, yeah, more. I just jumped to it.

Jake: I.

Co-host: Like, I wanted to know.

Jake: No, that was good. No, it's fine. I didn't really have more. No. Um, I. So not a ton. I did learn that schema does absolutely nothing for search engines. And, uh, most of the AI that we tested. So, like, there's an asterisk here. This is something I want to write in Search Engine Journal, because in Search Engine Journal, I just want to cover, like, The LLM side of things. And then in Moz, I would just want to do search engine side of things. But, um, if you open that report, if you just read the abstract and the introduction, that's really all you need to know what the report is about. And that's collectively 500 words. Maybe. Um, the rest of it is for. If you read those and you're like, yeah, but I don't believe it. And you want to audit it, that's fine too. But there's a really nice infographic up there. Um, so here's everything that schema does not affect. It doesn't affect Google Maps, it doesn't affect Google Search, it doesn't affect Google Mobile. It doesn't affect Bing, it doesn't affect Yahoo. It doesn't affect Google AI Mode. It doesn't affect Google AI Overview. It doesn't affect Gemini, it doesn't affect Grok. It does affect ChatGPT.

Co-host: Wow.

Jake: Yeah. So, um, I wish we could have done more with AI. I wish we could have like Perplexity in there and Claude in there. Um, but Local Falcon doesn't have those integrations. At least they didn't when we started the study and we kind of test for that. Um, so, but I will say this about Claude is that if ChatGPT considers it a factor or actually does look at it, there's a high likelihood that Claude does as well. Because Claude is literally a company or anthropic, I should say is literally a company made up of ex OpenAI employees. So you would assume that this is.

Co-host: I thought they were coming from Twitter. But you meant like former.

Jake: Yeah, former. Okay, so like they. It was literally like, um. Again, this is in Karen Howe's book Empire of AI. But um, she talks about how anthropic started because of a moral. Um. What was it like? Like a moral, like confliction between lead engineers and OpenAI and Sam Altman and so the lead engineers left and created Anthropic.

Co-host: Okay, um, do you know, have you like, what's the current state of the union for market share between these? I haven't looked. I don't even know. I remember seeing early on. But this has changed so quickly. Like is one dominating.

Jake: I just. Google. This is. This is Google AI overview. Uh, so of the market share for LLMs. Um, it's. It's split between so consumer. This is the consumer market share. You've chat GPT 60 to 68%. You have Gemini between 18 to 24% and then others underneath that, um, enterprise and API market share that's where you have Cloud 30 to 40%, chat GPT at 27%, Gemini at 20%. Um, I believe that a hundred percent. Because when I was thinking about the whole Claude thing too and how like, well, they probably use. If ChatGPT does, you know, uses schema, then, then Claude probably does. But then I thought about it for a second and I'm like, I don't, like, I don't think like consumers really use Claude to do research. I think Claude is more of like an automation thing. Like, like they're, they're trying to use CLAUDE for agents to do things. Whereas like ChatGPT and Gemini Focus, uh, more on like discovery and search, which is what SEO is all about.

Co-host: I only hear, I'm saying normies. I don't mean it insultingly. I just mean as like a, not us, uh, people living in this. I only Hear them mentioning ChatGPT. It's never the other ones. So that wouldn't surprise me.

Jake: Yeah, um, there is like a little asterisk too that I wanted to cover with like the Yahoo. Because in um, or uh, yeah, with Yahoo specifically. So yahoo had an 81.96% confidence level that, that chat or that um, schema markup did influence rankings. Um, two notes on this though. One, nobody uses Yahoo except the Japanese. And then two is 81.96% is lower than my threshold for marketing. Mine is 85%. Once it hits 85% then I'm like, okay, we should probably consider doing this. Um, medical and academic journals don't do anything less than 95%. But this is marketing where failure doesn't result in someone dying. Um, so take that for what it's worth. It's a little subjective there. I put in my, in my report, I said anything over 80%, like between 80% and 90% in confidence is a, uh, maybe. And then anything above 90% is yes. Anything below that is below 80% is no. Um, uh. And for what it's worth, the confidence test is determined by uh. So it's kind of a long word. It's a, uh, uh. Oh, two sampled Welch's one tailed T test. So, uh, you said that to me.

Co-host: I thought that was cool.

Jake: Yeah, it's very mathy, but, uh, honorary degree.

Co-host: That's what I'm saying. I mean, who do you want it from? Who do you. If somebody was going to give you a degree and said, jake, thank you

Jake: for your contributions to uh, Cornell.

Co-host: Okay.

Jake: Because, because I, uh, could make the uh, the Andy Dwyer references from the office. I Went to a little college called Cornell. Have heard of it?

Co-host: Nice.

Jake: Uh, no. I mean, I'd take one from you and I, My hometown. Yeah, that'd be cool. Hometown Hero Solves the Schema case. But no. Uh, well, it's a relatively good segue for what we're going to talk about today, which is AI. Ah, again. Uh, yeah, but only because I didn't. So episode, what was it? 208, I think. Uh, we talked about how the title of the episode was the Rise of the AI Informed Client. Which is. Which is Lisette's title for it, by the way.

Co-host: Yeah, she came up with that. That was banger.

Jake: So, um, she liked it because it was a Star wars reference and I appreciated that. Um, but in that, in that episode, what I wanted to actually talk about was how it kind of seems like agencies lately have been devalued in the eyes of clients or in potential clients. Right. Just in, like the general consumer market. Um, because now that people have access to these tools, it's brought the bare minimum up, uh, quite a bit. And now just everyday people and business owners have access to basically a graphic design intern. They have access to an intern for literally anything that they could ever want. Which means the, the barrier to hire an agency is, has now increased and the amount of clients that an agency, uh, is able to target has decreased because there's no longer a need for simple tasks.

Co-host: I think there's been. I can think of three things recently that I, uh, previously would have needed to rely on somebody with a graduate degree to pay for their services that I just figured out myself by chatting with Gemini. And that's crazy. Like, that's thousands of dollars that I just saved really, in my personal life. But stuff that didn't really have to do, but because I could, because it became accessible, that I wouldn't have to pay somebody a, uh, professional service fee. That's insane. And that just shows how much really you need to improve your service offering if you're going to continue doing well by offering services as an agency to people because they're not going to buy. Nobody's paying for entry, uh, level content writing. That's just dead. It's like dead as a market.

Jake: I can't even count the number of times that I've seen, I mean, in the past when someone said, like, in a Facebook group, like for, I don't know your niche, for example, like when I'm in landscape in lawn care business groups, I can't count the many. The amount of time that in the Past someone was like, hey, I'm looking for, or you know, a web design agency or something. And someone would comment something like, oh, you can just build it yourself. Oh my, my little light's not on. There it is. Um, that was important to me. That was good. And uh, and like, back then it was really easy, you know, to defend because you could just be like, um. What I would always say is anytime somebody was asking for marketing, ah, recommendations and anytime, like a landscaper, even better, like a lawn maintenance guy was like, oh, marketing is easy. You can just do it yourself. It's not that crazy. I would always comment and say, well, I have news about cutting lawns, so it was the perfect defense for it. But, um, you can still kind of use it, uh, because now that with, now that you have these kind of like AI builders or just like, I guess, quote unquote agents that can just do everything off of a simple prompt, I think it's really, uh, inflated people's egos and what they think they're capable of. And in a lot of, like, sense, they're not like, wrong, like, they are capable of more. But AI will only tell you the things that you are asking doesn't necessarily prevent future situations from happening that you aren't aware of when you're taking the action and consulting with it currently. And I have examples of it. But, uh, man, I, I just, I lost my train of thought because there's like a hundred different directions. I want to take this, um, but I asked, like, I'll say this. I, when I asked, I brought this up to Dan and Dan's episode. I think it was episode 206. Yeah. At 206, because Dan had ended up closing up his, uh, his graphic design shop after he hit about 2 million because of the economy kind of took a hit and he was coming back into the business and it just wasn't, you know, know, going as well. And I, uh, had asked him if he felt that AI was devaluing graphic design or just design agencies in general. And he, he, he said what you kind of said earlier, which was, um, yeah, but you have like, it's your job to like, illustrate the value. And if your value is trying to put people on like a simple retainer, then you're not as valuable. Uh, you have to find other ways to do that. And I think that's so much more true in today's age with AI and trying to illustrate the value. But that's like, exactly the problem though

Co-host: is that, but that's why I'm So big about hoarding data. People don't like it and they don't like when I say it and they don't want to hear it. But we can talk and pretend that we're in this new age where that is that important anymore and things are getting decentralized and everything is accessible, but it's not, right. Like, that is still the major massive thing that if you own and can use as a weapon in the best of ways, that's magical. Like that becomes the thing that is worth paying for is when people know how to use the data they've carefully curated, protected, analyzed and weaponized. I'm going to say it again. That's. It's awesome. And I like, that is an agency I want to work with is one that has done their research that deeply because that's a, that's a true differentiator, right. When, when people have done serious analysis like the schema report, right. That creates original stuff that actually gets put out, as opposed to when you just ask some generalized question to an LLM, it spits back out an answer based off of whatever it can go and dig up, which ironically you can influence if you are the person creating that information that then gets put out there to influence the LLMs. Right. Uh, so I don't know. I, I think that that's original research as a concept, right. I think gets, um. It's undervalued, right, by the agencies. Like you have the clients, the customers that are undervaluing the agencies, and maybe rightfully so at this point because they're offering services that they don't need to pay for anymore. But the thing that makes you valuable as an agency is going and doing that stuff, getting that advanced knowledge that you know and other people don't. And you can approach it different ways. I think publishing it is a great way to show your expertise. And then also, you know, um, it does help you get clients. Maybe not immediately. There might be like a few waves or people, you know, links removed from the, the person that you actually like. You know, we're not going to get clients from a schema study. Not really.

Jake: Right.

Co-host: I mean, it might happen in a few cases, but that by itself is not going to be the thing. It's going to be people who then attribute that with the company name and then realize, oh, they do this. And then it, it's a ripple effect from there.

Jake: Yeah, I think what you're hitting on too is like the, the, the establishing authority defense which I had in my notes, which is like we, we spent Last episode, basically an hour of us just complaining with no real solution, but it was complaining about our problems. And so like, I know, like this episode, I didn't want it, I didn't want it to turn into like a, just a complaining fest. And so like, I wanted to provide like, actionable things that we're, we're doing in our agency to kind of combat this Dunning Kruger narrative. Not even Dunning Kruger narrative, Dunning Kruger effect on like, uh. So I was talking to a client today, um, and it was uh, it was the uh, I'll just say the first letter, the first word, elite. And they were, um, uh, I was talking to him about uh, he was saying how he just had an upset client because, um, they've been using uh, chat. The customers have been using ChatGPT to create like landscape designs and then give those to landscaping companies to be like, hey, I want a design like this. And then that way can help them kind of conceptualize that. But, um, he said he got uh, like in trouble from a customer because the customer wanted, um, because he sent over a design that was used with his actual design software. And they were upset because, you're going to laugh. Chat GPT knows everything and therefore the designs are better.

Co-host: Wait, what?

Jake: Uh, the customer was upset because they did not use a design generated by Chat GPT. They use their own design from their own design software from their own brains.

Co-host: Oh, and they're saying that what they were able to do, do with Chat GPT is superior.

Jake: They didn't even insinuate that. They just said Chat GPT is better because they know everything. It's like one of the things in the design process and one of the things in the design process is that it lists out like the plants and like the different things that you should use. Not just like. It's not just a visual representation, it's also like a descriptive representation of like, what is actually going to go where these um, you know, things are. And our contact was saying, he goes, I, he goes, you can give Chad GPT. Like I've tried to do this, uh, I've done this a few times. Use Chat GPT just to test it. And he's like, and you can give him like the, the sun and shade and soil temperature conditions of things, but it'll still recommend plants that are just going to die. Um, and so, so he goes, it's not superior. It's like it's trained off of, off of what everybody else publishes out there, which is nine times outta ten wrong. Or there just is a lack of information on the topic. And so I was telling him the same thing about how, like, what he's going through with his customers is no different than what we're going through with ours, because our clients, like, we call them customers, you can call them clients, but, like, our customers will come to us and they will say things like, hey, I ran a, uh, website audit through, through Claude, or through Perplexity or whatever, and these are the things that it said. I just want to make sure that we're working on these things. And that frustrates me so much. Instead of like, hey, we ran this audit through, through Claude, um, uh, can you explain more of these concepts? That's totally fine. Sure. You're trying to learn SEO. It's something. You're learning these new concepts. I can explain this. But, like, when you come and say, you know, I ran this through Claude, I, um, want to make sure we're doing these things. And I know, I know we talked about this in episode 206, but, like, you're, you're coming at me with the idea that you know more about SEO than I do because the, because you learned everything from the AI you're consulting with, and that is by definition the Dunning Kruger effect. Um, you don't know what you don't know. But anyways, um, you mentioned something. You're like, uh, earlier. You're like, it's, it's. You can influence it, and you absolutely can. And this is the whole reason why we wrote the Martech article, uh, uh, about why it isn't a red flag for agencies to own ad accounts, because we own our client ad accounts because of multiple operational, legal, and insurance logistic reasons. And every single article out there on the Internet talk is so any, any time that, like, ChatGPT recommended this, it would, it would link to a source which was just an article written by another agency, and it was that agency's sales pitch to, uh, to. To get you to value them over somebody else. And that's all it was. So I wanted to publish an authoritative piece where that information was not publicly available to, uh, LLMs on the other side, or I guess on the same side of the things of things. Uh, if you ask Claude or Perplexity or chatgpt, what are SEO factors that you should be focusing on, it will undoubtedly bring up schema markup. And schema markup is one of those things that's just complicated enough for novice SEOs and agencies to promote the hell out of, because it is just complicated enough to confuse and disorient clients into thinking that they need to pay more for the service when they're not getting the actual results that they should be doing when their header structure isn't even aligned. So because of that and there's a ton of information out there about why schema is important written by these same types of agencies. So I wanted to create the most authoritative bulletproof case study on how schema is not that important. And so that's why we conducted this test. That's basically my hypothesis on it. And the hypothesis was proven true in like what, like 10 out of 11 categories. Um, and uh, the data is all there to show it. Now this, this goes into like the credibility and the authoritative defense that we're putting in um, ahead of time now the amount of work that goes into it is just like unfathomable and I am exhausted. Literally like I could shut my eyes and go to bed right now.

Co-host: Yeah, um,

Jake: I don't know.

Co-host: I, that it's worth it, it's worth taking time to do that stuff because what is the alternative at this point? You really. I uh, think that if you don't do that and if you're not moving in a position to do that, I also think people way overestimate what that means.

Jake: Right.

Co-host: Getting to a position of being a thought leader in any micro niche, it's not that crazy. Just start working, just do it like just start moving in that direction and you'll gradually build up over time and like if you're scared of the time, it's just not the right fit for you because like Jake says the time's going to pass anyway. So do it. That's the future. Like that really is uh, at this point. What else do people have to go off of when doing these source sorts of, you know, analysis. They're looking at two different agencies from each other. They seemingly look, you know, the same. They offer similar services, they're going to get sales pitches from both. But again the floor is raised because they already probably doing a decent amount in house at this point that they've self taught themselves. So how are they going to decide between the two? It's probably the person who has said more and actually has information out there showing that they're doing serious original work. Because if you're just copying what other people are saying it doesn't work anymore. Like it's dying. I think that is what is devaluing.

Jake: Um, that's the other argument too is like, is the like LLMs are always going to be trained off of Their most recent training data, which could be, which is usually years old because they don't do training, uh, every year, every month or whatever. It's like, I think last time I read it was like every three years. They plan on doing it because the amount it costs to train an LLM on things is insane. So that's why you see um, not even LLMs but AI in general. You see very telltale signs that a site is built with AI. So we recently just had a client that rebuilt their website without telling us and they used completely AI. It was a complete AI build. They even moved like the, they updated their A record to point away from our server to their server. And um, when I looked at their site immediately I knew I was like, this is AI because there's ah, a rounded callout, uh, tags that AI absolutely loved. But the problem was I wasn't familiar with the AI they use or the tool they used. So I plugged it into the code and everything into Gemini. I was like, hey, what is this built with? And um, it told me the app and I'm like, it looks like completely AI built. And Gemini was even saying like, yeah it does, there's some telltale signs. And it showed the things that I agreed with that I already had in my mind that looked AI and it pointed out the exact same things. And not saying that Gemini was 100% saying that, yeah, it is AI because I went in and looked at the code and I gave my AI brother a handshake. No, but I already had these ideas of what it was and Gemini just confirmed it. But um, uh, regardless, it's a real

Co-host: strange place at the moment. I don't know how to feel about

Jake: it because I remember looking at that situation specifically.

Co-host: Yeah, just AI websites that's happening more and more and uh, they're micro sites whenever we see them. Right. I don't think anybody's building absolutely massive sites, but the simple five page site concept, they're getting spun up relatively quickly and they're not awful. Um, but they don't work. That's kind of the biggest thing is they have, I don't know, 90% of it pretty solid and looks good on the surface. And then when you start poking around is when you realize that important things are actually broken and that's the hard thing to fix.

Jake: Um, I think the biggest problem that people aren't looking forward to is like future integrations and changes for this. Unless your AI builds you a CMS that you can edit every single piece of text, then it's kind of worthless to me because then you have to prompt it to then change it. And maybe there's tools out there that do that really easily. And I'm just unfamiliar with it because I do things the old school way. Um, I gotta refrain from saying things like that because after episode, after one of our episodes came out, I think it was like episode 207. Oh, uh, no. Yeah. Was it 207? AJ was like, this episode should be titled, uh, Old Men complain about new technology and Refuse to change or something. I was like, aj, that hurt us. Like, you're not wrong. Uh, yeah. But, uh, like, case in point, when my students at the high school were building like an MVP for their website, they used replit to, to build that, but they couldn't change like a single line of text because they needed to re prompt replit to do it. They were out of credits. They couldn't do it. Uh, but having said that, even if they had credits, like, I wouldn't want to reprompt to change things. I just want to go in and change it because I don't want it to mess with anything else. Um, I'm okay.

Co-host: Like I'm. I don't know, I want to say out loud, I'm okay with it. This is where things are going. It's fine. It's going to eventually work. They're going to figure it out. And then a lot of this. I don't know the timeline and I don't think it's worth. Lots of people are going to make a lot more money than me by pretending that they know how quickly these things are going to happen. They're going to end up in news articles and charge people lots of money to go and talk about it as if they actually know. And, uh, nobody does. Perfectly. Right. But it will happen. Uh, uh, probably even at this point, the concept of being a dev. Right. Will, uh, I mean, it's going to change absolutely dramatically. Most likely people aren't going to write code at all.

Jake: Right.

Co-host: It's just going to be cleaning things up or adjusting things. And that's a crazy thing to think. Um, but at the same time, what, so what, what does that do for you in this current moment?

Jake: Right.

Co-host: It's like you can vibe code things and that's. It's just like a fundamentally different business model. I've been thinking about this one a lot lately because I have been thinking about next year in the business. Business models that are appropriate things that make sense for people to actually pay us for that they're going to get value in exchange for what they're, they're paying. And then in addition something that would be hopefully internationally scalable in some way. And it's, it's crazy hard to do at the moment just because of how quickly these things are progressing. But I don't know, I think. Who was it that we had on? It was um, Stark, Jonathan, Tony. Not Tony.

Jake: Yeah, Jonathan from. Jonathan, uh, Stark from uh, uh. Oh man. The title of his book is why

Co-host: Hourly Billing is Nuts.

Jake: Hourly billing is nuts. That's it. Yeah, yeah.

Co-host: Um, it doesn't work for everybody. I think the idea of being everybody dreams and wants to be in the top, you know, 1% of whatever skill that is. I used to think that too. I don't need to be that high. I'll take the top 20 is cool. Uh, as long as it's, you know, noticeably better than the majority of the others. Uh, that is the differentiator. That's just what I keep going back to. I said it doesn't just have to be original research by itself. There's tons of ways to do it. And actually I do feel like an old man because some of them I just don't want to do because I do feel like I'm from the old Internet. Like I like, I like writing. I would prefer to write new Internet is video. That's just how it is. So it kind of values and same thing that happened to uh. This is like politicians back when the TV was made, right. You didn't have to be good looking until there was a tv. Now you have to be a good looking politician. Before that all you had to do was talk. Well, because it was radio. So that's how I feel.

Jake: What uh, what election was that? That was it uh, was it the JFK elect? No, I think jfk, was it jfk.

Co-host: That's right.

Jake: I don't know because I remember, I remember talking about this and like they were going out there and they were going to do like hair and makeup and one of the candidates was like I don't need to do hair and makeup, I'm a man or something like that. And like JFK did it and like he pulled better on TV versus than on radio.

Co-host: Yeah, that's where we're at when it comes to influencing online too. I mean that's how people are judging agencies.

Jake: Yeah.

Co-host: Their type of content that's being put out.

Jake: I think like the, like the biggest defense too for, for doing this is like you're saying is like getting that content out there whether it's like research like what we're doing or whether it's through what like um, Caesar is doing with TikTok and his like just social media presence. People want to work with Caesar because he's, he's a person that they relate to on, on social. People end up wanting to work with us because we know what we're talking about now. That doesn't change. You know, six months down the road where they're working with us and they're past the honeymoon phase and now they're question that were doing because they went to some kind of seminar, uh, or webinar and uh, were encouraged to Explore SEO through ChatGPT. Um, and I think we're always going to battle that. But I think the hardest part with some of these things that people now think they're capable of is when they get to a time in their life or their business where maintenance needs to be done on the thing that they were consulted to do with, on like ChatGPT or something or like something needs to change. So like a really good example is. Okay, cool. I built an, I built an AI website using replit. Great. Well did you know that is client side rendering and it's not even crawlable by search engines. So you know, maybe you should have thought of that. Maybe you should have had repl.it build you a server side rendering, uh, website. Oh, you didn't know about that? Well, why didn't you just ask the AI to build you an SEO friendly site with uh, those capabilities? And um, it's because it's again, it's Dunning Kruger. Like you don't know what you don't know and everything you do know, you feel like you know everything about it.

Co-host: Well that's, that's like the fundamental difference in how people are using AI and how they're like. I would argue, I just say that you're doing it incorrectly if what you're doing, if your angle, if your primary angle about approaching AI at this moment is using it the same way anybody else would in order to accomplish the same low level tasks. I just don't believe that's valuable. I really don't. And I think it's moving so quickly that by the time you actually do get any momentum, it can't possibly keep going because what value do you have? You're doing nothing really helpful just by offering that. But what the massive value is how quickly at this point in time, AI can accelerate your learning for something that is genuinely complicated. Right?

Jake: Yeah.

Co-host: The number of things that talking about that server side rendering, just that concept about what's, what's loading and where, like, you can go. Learn that now in a way that would have taken you a whole lot longer to go down a rabbit hole and figure that out. Like, you can just have a conversation and they'll teach you. And you have to be careful and you have to like, put up your own guardrails so that the conversation doesn't just go in the direction in which you kind of want it to go. But if you do guide it objectively, you can learn things and you can, like, that's the crazy thing. You can be an expert in something that used to take way longer just because you can teach yourself. And that is the. And then, like, then take that and use that part with the AI and sell that. That's fine. But like, send emails. Yeah.

Jake: And it still says wrong things, like, even when it's explaining a topic. And like, I've, I've recently, I've had to like cross check it. So like, I'm creating a. So we're doing like a ten man paintball tournament in the Iowa City or University of Iowa Connecticut stadium. You're invited. You'll be a ref at that, at that event. That's gonna be cool. That's history, dude. How do you feel you're making history? First paintball tournament on a college stadium.

Co-host: I don't know how to ref, so I gotta watch a YouTube video or something.

Jake: But I can. I mean, he's only gotta know a couple things. It's really easy. Um, um, but in order to make this event go super smoothly, I had Gemini, uh, create like a round. Like, I have five teams, a Gemini create like a round rabbit tournament style. And then there's a runner, which is me. And the runner goes to the press box to get the next team and brings into the sidelines, takes the sideline team back to the press box, all this stuff. Um, so to make sure we fit within the two hour gap, I was like, hey, I need you to like print out like a schedule for me. Like, that is like, I need to take this team back. This team's on deck. These two are playing, these are rotating. Um, and it sent it back to me and it was like, correct. But then I sent. Then I said, hey, I want to know how much time each team gets in the press box in between, like in their breaks. Because I don't want them to go to the press box for two minutes and then come back. And then, you know, it's like, well, why did we Even go, um. And uh, it sent it back and I was reviewing it and just like looking through. I could have just printed it off and called it good to go. But I was reviewing it and it was like not even close to correct. Like it was just like it looked right when I looked at it like initially. And then I was just doing. I was like I should probably check the AI because it's probably wrong. And I was like it didn't even need to like research information on this. It just needed to be like logically correct and it just wasn't. And so like um, everything that I learned now from an LLM I have to go and actually like I either ask it, can you, can you like you're saying this as a fact and you like source it, like give me the article that you're pulling this information from or an article you're pulling this information from. And I don't think consumers are doing that like customers and clients. I think that a lot of them are just. I mean this example with our, with our uh, client that, that got in trouble for my customer because the customer think that thinks that Chad GPT knows better than, than him the expert. Um, yeah, I don't know if that's not Dunning Krueger, I don't know what is.

Co-host: Yeah, it's so, I don't know. I think we, especially you like saying we but I'm self injecting and including myself in your communication. But I've seen you proactively communicating it to clients about how we're approaching this and being constructive in ways that you can help them. That's another weird thing, is just helping normies approach this stuff in a way that's helpful and practical instead of, you know, doing crazy stuff like falling in love with AI. Uh, like I say that jokingly but you know, like that's a real thing. It's weird right? Like the, the things that people choose to believe when they're allowed to and not guided by like external, you know, any other third parties that can be more objective and they just kind of instead let themselves go down these rabbit holes and then we end up. Yeah. Having those conversations. And that's. There's value in that. That's part of the value. And being a uh, true strategic partner is being honest and telling those people those things instead of just saying may I take your Hassler? This is winning at juniors.

Jake: The problem is I don't think these conversations are avoidable no matter how much authority or credibility you put out there. I think people are Always going to have these conversations of trying to prove your value. I think I said it in this podcast in the last episode, which is someone asked me what I thought about how AI, something about SEO. And I was like, I don't trust AI SEO advice. Because xyz. And they're like, well, you know. Yeah, yeah, I get it. I think. Oh, I remember it was our Iowa client. And he's like, yeah, I know, I understand having to, you know, you, uh, know, defend your, your business against AI. I'm like, that's not what I'm doing. Like, I'm not defending my business against AI. Like, I'm just saying the autistic truth, man, like, like unfiltered. This is wrong. And you should be ashamed, uh, for believing this. And like, um, I. So, like, having said that, the worst part too is that these conversations are totally unavoidable. And no matter how you try to frame things as like, hey, this is wrong, like that what AI is saying, customers and clients who still put their value in the LLM or in AI will see your opinion as bias because you're trying to defend yourself against the AI. And it's super unfortunate. Um, and I think that's where everything just needs to be relationship focused. People should just like working with you. Um, and that's probably also kind of ties into maybe the product ties model a little bit in the episode that we just recorded, uh, well, recorded yesterday, but it came out last week about how like, um, maybe like that's kind of an emotional drain. It's an emotional drain where, like, we. The relationships that we're having with all these clients, as they start to scale up, it gets more and more impersonal. And with the productization and because it's impersonal, like we're really no different than ChatGPT or Gemini to these clients.

Co-host: This makes you realize just how personalized, like, career and professional advice should be to. It's not age group, but it's experience levels. Right? Because relatively early career hard skills are very important. Like developing hard skills that are. You have a marketable skill that's worth money that you can sell and people want to pay you for. Uh, and that gets you kind of the baseline amount that you can earn from whatever that skill is. Maybe you can earn a lot more

Jake: if you're better, right?

Co-host: If you're objectively better at it than other people doing that hard thing, then that's one way in which you can receive more money. And then you realize that there are people out there charging 2 times 3 times the price that you or somebody else is charging for no other reason than they are a class act to work with. Right? Like, that's. That's how they do it.

Jake: I have one anecdote for one story, I guess. So I sent this, uh, remember, uh, in March, Ezra, uh, one of our friendly competitors forwarded us a lead for someone who was trying to set up an API connection between their form and their CRM. And they're trying to bring in UTM parameters. And they worked with two previous agencies who couldn't figure it out. And so I was like, yeah, sure, I can do it. I can AI my way through this. I consulted with the Gemini. Gemini was wrong in multiple situations, and I ended up doing the opposite of what Gemini was saying. And it worked. Um, but Gemini just having that thing to talk to led me through the answer because I could use my own brain at the same time. Um, however, talking with this contact, uh, I was. She. She asked me if she could, like, pay us to add schema to their site or fix their schema on their site. And we were like, this is March. So we were like, midway through this, through this, like, research study. So I had to give her my honest thoughts about schema. And I'm like, look, um, like, schema's pretty easy. I could just update this for you. It's. It's not hard to do. However, I don't think it's something you should be really concerned about because. And then I told her, and I said, like, look, if I honestly believe schema was an important factor or something, I want to be having this conversation with you. I would just be sending you a bill and then we'd be doing it, um, like that, that's. That's what the nature of that kind of relationship is. But that's not who we are. So she goes, hey, when this schema report comes out, I want to read it. Send it over to me. So I just sent it over to Day and she wrote me this like, huge, long thing. Like, I can tell she read every single word.

Co-host: That makes. That makes one so far.

Jake: Yeah, I don't, I don't think she's going to be a client, and I don't really care about that. But, like, uh, like the fact that she has pull in the industry as well. She's the one that has like a hundred Poop Scoop locations, uh, in that, in that business model, which, frankly, I don't care about getting into the poop skipping industry any longer. Um, but we just can't figure it out. I don't know what we're doing wrong. It's just.

Co-host: It's a thin business.

Jake: It's a thin margin business. And also it's not one of those things that people really search for. You have to be really good on platforms like Facebook and Nextdoor with offers and such. Um, and we're not positioned for that yet.

Co-host: I just don't. I feel like if you're financially and operationally oriented, as you progress in business and you see two different paths, you have working with more people who are continually running thinner and thinner business models. And by thinner I mean easy to fall apart. The other path, which is more complicated and whatever. Like, I don't really care what the barrier is. It can be financial, it can be professional, like skills. It can be some, like, certificates, whatever it is.

Jake: Right.

Co-host: Um, I want to go that way. I definitely want to go that way because I know that that's way more stable. Like, that's I don't have to worry about. Here's what I can say. Um, I don't know. People making microchips or semiconductors, I don't even know how that works. I don't know anything about that. But if I'm betting money on one person going a little crazy, drinking too much at the bar one night and deciding that they're just gonna walk away from their business and go, I don't know, sell exotic animals or something, something truly crazy, I don't. I will bet money that it is not the semiconductor guy. Right. It's going to be the thin business. Something that somebody can set up overnight or, you know, with less than 5,000 bucks. Startup Capital, website, Boom, Pow. That's kind of done. Like, that's just operationally. That's more of the direction to go in, I think long term.

Jake: Yeah. Yeah, that makes sense. Well, I don't have much more to add to this. I just wanted to come in here and talk about how, um, I've been noticing a trend. I think I'll end it with this because I wanted to say this. I didn't say in my notes, but I didn't say it in this episode, but I wrote in my notes and it's that, um, uh, in ways that agencies are being devalued is. I did say this though, but like, we don't need a dev agency to set up a simple API anymore. We can just do that. Uh, we can consult our way through with Gemini now. If it breaks, maybe. Yeah. And that's the argument that devs are going to make. Um, is that if it breaks or if there's other complicated nuances that you can't figure out or something is deprecated and AI thinks it's not, um, that can cause issues too. But for nine times out of 10 applications, we could just do it. Now. We, um, don't need a designer to make Facebook ad like simple Facebook ads. And if we don't need that, then customers don't need that either. Um, they can, they can do it on their own now. Now if you want to like a B test different things and, and change one little image, one little thing in a, uh, in an image, then you know, maybe that's, that's the case for designer. But also if, if customers get access to Canva and Canva AI and different, and create multiple different assets, they do have that power to create those individual, to change those individual elements and do the A B testing. So it's only a matter of time between when those values start to, um, diminish even more. And then where does your value as a designer or a dev lie? And personally, I think it's going to come into play in performance and consultative work. But if you don't believe that, for example, you don't necessarily believe that your value is being diminished by AI because you're this high and mighty dev and designer. Like, I think we're excellent. I think we're amazing at what we do with, with SEO and Google Ads. I don't think anybody in the industry can beat us. But that doesn't mean that I don't think our value is being diminished by, with what AI can do or at least with what AI is promoting itself that it can do. Um, it's all about what's happening in the consumer's mind, not about what's actually real. Um, the last thing I was going to say is like, if you don't believe that, I mean, just look at the industries that maybe you have consulted with AI AI on like LLMs, like, um, law. Like if you ever had a question about legal advice or not advice, but just like the law in general, I bet you didn't go to your lawyer first, that you probably went to Gemini or ChatGPT or other stuff that I've

Co-host: done without a lawyer at this point would scare any lawyer. For sure. For sure. And it has worked out.

Jake: How much. Wait, how much time have you spent with, with Gemini or Chad GPT on the trademark stuff?

Co-host: Um, hours. I mean, many hours.

Jake: And then also, I won't say thousands. Yeah, I won't say who yet either. But you also. You also took a really big company to court, Like, a big company to court, and you've already received judgment on it.

Co-host: Yeah, that happened.

Jake: And you consulted with Gemini the entire time.

Co-host: I didn't know what I was doing. I didn't. And, uh, yeah, it just walked me through the whole thing. It's crazy. It's real insanity. The things that. You know, that there's the bell curve of life and people's intelligence and that you want to think, well, I mean, somewhere in the middle, and then people are smarter and people are dumber. And then I. You know, in Home Alone, when he uses the credit card and he's like, wow, it worked like that. It's like, that's how I feel with what I've been able to do with some of this. At this point, I'm just like, that should not have worked. And it did. And I'm concerned about the people who are on the receiving end of it. Maybe they aren't. Maybe they're not people. Maybe they're AI too. It's not actually going through a human person, because I can't imagine that. I don't know just what I've been able to do legally by using Gemini as my lawyer.

Jake: Now, granted, obviously, we're not advocating to use AI for every legal situation.

Co-host: Yeah, no, they'll don't do that. But, like, there's a time when you want a real lawyer first.

Jake: But be honest, though, like, in the broad sense of the word lawyer and the things that you typically thought you needed a lawyer for, has the value of a lawyer decreased for those things?

Co-host: Crazy. It's crazy how much it has.

Jake: Yeah, yeah. Uh, and the other thing is, like, finance, like, how much. Have you consulted with Gemini on this tax situation in Japan?

Co-host: Yeah, that's okay. You're hitting some real home runs right now.

Jake: Okay. When you had your back issues, how long. How much did you consult with Gemini when you. When the Iowa health system failed you?

Co-host: Yeah, this is like. This is what I was saying. Uh, because you had just hit, like, three graduate degrees right there. That's the crazy thing is we're not talking about, uh, I'll put quotes standard, college educated.

Jake: Right.

Co-host: You have a bachelor degree. Something like that, which would normally, historically, people would be like, well, you're locked in. You're solid. That's great. Um, even finance, like, that's generally, you know, people would put that in that category, um, because, like, ah, an accountant depends how far you're going. But yeah, um, that would just be a bachelor degree sort of Thing, like all of those. All of those. Uh, and it's because for me, this is probably just me. Like, personally, I think the crazy thing is, for me, it's not the, uh, entry level that it's wiping out, because a lot of these things that I'm doing are relatively complicated at this point. Like the tax thing with Japan, that's. That's insanely complicated that you're talking about the absolute top level of expertise, like somebody who has spent 30 years building up their, you know, reputation and professional expertise. And, um, to be fair, if I do something like that myself, that doesn't mean that everybody else. Because that's like our. That's kind of our point. That was my point earlier, too, is when you put yourself in that position, you're going to, like, some people are always DIYers. They're always going to do that, but then the other people who aren't are going to be looking for that person in authority, in that position. I am not the person you want to compare to in that situation, because I look at that and I go, oh, I don't need to pay that premium for you. Yeah, I don't. Um, at least not yet.

Jake: That's not most people, you know, because actually.

Co-host: Okay, so to be fair, I was emailing with, um, with a lawyer where I. I just told them, like, hey, um, probably because I could use, like, some consulting at first, at the beginning here, but, um, I'll. I'll probably want to do this myself, granted, if my income trajectory continues the way that it is. I don't want to do this myself. Like, I do want to pay somebody to do this. And, and that's like the dream client for any agency is, I know I can do this myself, but I don't want to because I want to pay somebody else to do this, whose job it is to do this dream client.

Jake: Right. Yeah. And at least not yet, too, because, like you said, like, you're consulting with the finance, the financial situation with Japan. Um, and you're getting to a point where, like, you're being armed with the information that you need to be armed with to go find the person you need to find.

Co-host: Yeah.

Jake: And. And like, that's what's important. But, like, you're not going through these, like, initial steps of consultations of wasting money with all these people who, like, can get you to the next step. You're just skipping all those and going. Going to that. So literally in this conversation, we have devalued legal, medical, and finance. And like, and like, so imagine like you're a web designer or you're a graphic designer. Like, now every single one of those can make the argument, and rightfully so, that that is not the case, that they are still valued at the. At the highest end of things, which is true. But in the mind of the consumer and the average person, you have to admit that they value you less. And so now you have to go on the offense on how you're going to portray your value ahead of time. That's all.

Co-host: That's good.

Jake: All right, we'll end on the good note. So, everybody, if you enjoyed this, I, uh, promise we're going to stop talking about AI really soon here, because I'm getting tired of this. But if you like, it makes me sick. Um, these aren't even the episodes that do well. It's just, like. It's just what's happening right now. And. And, like, we're not. We're not catering to what people want. Nobody's asking us for this. So, like, if you want to hear something, let us know. We'll talk about it. Uh, yeah, I did the website. We got sent a guest lineup while we were talking. Two different things there. Yeah, send us a message. We also have a guest lineup that is, uh, we have more guests coming on. Um, and, uh, we just took a break from that because the agency was just kind of, like, underwater. We need to take care of that first. But, um, yeah, also, in. In accordance or in. After you send us a message or before, leave us a review on Spotify or Apple. Uh, and if it's negative, please email us after you leave the negative review and tell us why. That's all.

Co-host: Yeah, we want to know. Tell us how we can improve.

Jake: Yeah, so, yeah, make a burner email. I don't care. I just want to know why. Or dumb.

Co-host: We're not going to. I mean, I might be hurt. I don't know if you're going to say really mean things that have substantial, like, actual substance. We're reasonable people. Right. So if you're just like, I don't like you because you make something up. Uh, like, we're not gonna be hurt that way. But if you, like, have a really actual thing, like, I don't know, I'm not gonna give anything because then you would say it, and then that would hurt. So I'm not gonna do that. But anyway. Yeah, I'll just stop here.

Jake: All right, everybody, thanks for listening. We'll see you next time.

Co-host: See you.