#589 Why AI Search Will Break Traditional SEO (And What Actually Works Now) with Joe Toscano

Mehmet: [00:00:00] Hello and welcome back to an episode of the CT O Show in Mehmet today. Are very pleased. Joining me, Joe Toscano. He's the founder of Service Stories. Um, Joe. I don't like to steal much of my guest time when they do the intros, right? So I'll leave that to you. But what I like to always to give hints to, to the audience, what we're gonna discuss about today.

So we're gonna discuss about, you know, the shift in marketing, specifically when it comes to search engine optimization, the shift to the AI world, how businesses they should act on this, and from operational model also as well. So without further ado. Um, tell us more about you, your background, your journey, and then we can start from there.

So the floor is yours. 

Joe: Yeah. Well, first, thanks for having me. Uh, it's always great to be on these and get to know somebody new and talk to a different audience, so thank you for having me and giving me that opportunity. Um, my background for those of you who haven't heard my name or uh, [00:01:00] seen me online is.

In technology. I have been a product designer and developer for now going on 15 years. Um, but if you looked me up, you'd see that I spent about the last 10 years fighting for a safer internet. Uh, that happened because I built up a career very young. To where I was consulting for Google out in Mountain View, and as I was there, I got concerned about the way data was being used in the Valley, not just at Google.

Um, and I wrote a book called Automating Humanity, which is all about the premise that. We are trying to automate everything that is human, but yet mm-hmm we can still make it humane. So that book took me on a world tour. Uh, I have helped shape laws around the world focus on data privacy, especially protection of kids.

Um. It also got me featured in a film called The Social Dilemma that you may have seen on Netflix. It's the most watched technology documentary in the [00:02:00] history of the world as far as we've been told. And uh, I've done a lot of impact work because of that. But at the same time, I was also. Asked to help my parents do the marketing for their auto shop.

So imagine I'm living this dual-purpose life where I'm flying all around the world in these, in these big rooms, uh, helping shape laws and do these things at the same time, like helping with this local auto shop and, uh, digitally transforming it and helping with their marketing, all that. So it, it all plugged in really well because ultimately at Google, you know, I helped with Google my business, which is now called Google Business Profiles.

Um. As well as Google AdWords and the merger back when they didn't have AdWords and ad, you know, Promax and all that stuff. It was merger AdWords Express was the thing back then. 

Mehmet: Mm-hmm. 

Joe: Um, and yeah, I learned a lot and applied it all to my parents' business. It grew four x in five years and, uh, basically I knew the next phase was going to be getting them into AI search.

[00:03:00] Um, I always knew it was coming 'cause I saw it at Google, you know, 10 years ago. Uh, and. That was my big exploration when I moved back home to Omaha, Nebraska. And so what we can talk about today and what I'm really excited to talk about is kind of how that happened. I'm happy to dig into the details of that transition or the, the learning.

Sure. But also, um. You know, the actual application of it, because AI search is different, but it's not like collapse your web presence different. I don't like to frame it that way. I think it will be if you don't adapt over the next couple years. But I think people are getting sold fear right now that it's, it's gonna just rip them out.

Uh, today. Um, I see it more as like. The difference when we went from desktop computing to mobile computing and Google changed our algorithm. So if your website didn't, uh, perform well on a phone, you started to rank horse. I think the same thing's gonna happen if you don't create content in a certain way.

If you don't build your [00:04:00] website with certain parts to it, um, I do believe you'll start to rank worse over time. 

Mehmet: Joe, thank you very much for this, uh, very rich introduction. And this is why I like to keep it to my, to my guests to do that. Uh, no one can tell the story better than the ones who lived it. And, um, and, and it's like really fantastic, you know, spending the years at Google being featured in the social dilemma.

I watched the v the, the, the movie by the way. Um, and, you know, telling your story now, just, you know, a question that came to my mind because you, you, you talked about, you know. How things keep changing. And this is like the normal in, in technology. Like 

Joe: yeah, 

Mehmet: you just gave the example of moving to a more mobile world and how you needed your website to render for the mobile and also like, you know, you needed to appear in Google search and all this.

So now, 

Joe: yeah, 

Mehmet: AI is doing another shift, let's call it this way. Um. What do you [00:05:00] think like is the biggest underestimation that is happening between like business owners, founders, operators, that, you know, they, they're not getting what AI really is doing here? Because you mentioned also they're being sold like on fear, which is we see it a lot, but really, what, what's the main underestimation here?

Joe: Sure. Well, I think there's two ways to approach that question. One, like what is the main thing today? And then also what is the main thing maybe gonna be in the future? Um, because I do see AI changing our world radically over the next five and 10 years. Um, but immediately I think the thing that people are missing is the.

Ways to communicate with ai. And that sounds really like we're talking to aliens, but, um, the fact of the matter is, you've always built your website or whatever you're working on to speak to Google search bots, right. And now. We're having a different [00:06:00] conversation with ai, and it's similar in the early days, right?

Like the, the technology's not so far different that it's like a radical advancement. But one thing that I'm seeing and that people can go look up is the, the sense of a query fan out, right? So this is, uh, a new thing with AI where if you put in auto repair near me, it's pretty obvious what you want.

Right, and so Google can go find those keywords or things very close, much easier. I consider that era of Google search and SEO like spear phishing, right? You're shooting like a straight shot. You're trying to target for these explicit keywords. Today someone's picking up their phone and having a conversation with ai.

They're going. I just ran into the curb because I slipped on ice. My, my truck hit the curb, and now my wheel's bent like, what does this mean? What do I need to do? Click ask ai, right? Like, that's what they're doing. And, and [00:07:00] so Google and Chat, g PT, and Claude, they're then gonna take that, you know, paragraph worth the text and they're gonna try to think of five or 10 or maybe 20 different queries that could, uh, uh, um, properly feed into that and feed them the right information.

So it's kind of, you know, it's a transform. From the big paragraph, human text to, you know, these query sorts. And then they're gonna go look through that. And instead of you being the one result for the one specific keyword or phrase, now maybe you wanna be in five or 10 different phrases, right? And you wanna be actually like number one on this one, and number two on that one, and number three on that one.

Because if you're number eight on this, and number six on that, number one on this, maybe someone else is higher up. So it's a. It's a different playing field. I think the playing field's a lot bigger now. Um, that was something I learned when I was, I was programming chatbots back in the day, you know, teaching it the difference.

Like [00:08:00] what does hi mean? Is that Hi, hello? Or is that hi like Hawaii abbreviation, right? Like it so basic, but you're like, we, that's what we had to teach bots 10 years ago. Um, and so it's the same thing now. Uh, the playing field is much. Bigger, I believe. And so there's a lot of opportunity to to grab land while it's still early.

Just like the early days of SEO, right? You could literally put up a blog posts for any specific keywords and you would just skyrocket to the top 'cause no one else did anything like that, right? I think the same days today. 

Mehmet: Joe, I gonna ask you something, which it's a follow up question, which I didn't prepare.

Um, oh, oh, it's related to the question I wanted to ask, but I gonna reframe it in, in, in a different way. 

Joe: Yeah. 

Mehmet: Now to your point, and this is something, even myself, I experience it like a lot of my guests, they said they experience it the way we use. Actually, you know, we, we use. So first we use chatbots like chat, GPT or cloud.

Joe: Yeah. 

Mehmet: [00:09:00] Similar way to how we used, um, Google for search. Mm-hmm. And then we figured out that there's this prompt engineering thing and you just like put things out there. 

Joe: Yeah. 

Mehmet: Now. Some people you know, and then you, you have the search within, you know, these chat botts so they can go and search. So someone might say, Hey, like, is it really a shift happening there to SEO?

Because actually the bot that chat, GPT or like cloud, they're gonna go and do a query inside Google or I, I don't know if they use anything else, but I'm sure, I'm sure the measure that they still use Google and they gonna. Be, you know, applying the same methodologies of search engine optimization to that.

Do you agree on this or no? Like, you know, the, the, the way these, you know, AI search is completely different than the way you do a search in a search engine like Google. And the reason I'm asking [00:10:00] you this, like how much this will affect the visibility, because you mentioned like it's a spearfishing, like it's you, you, you just tried.

I need a, you know. I don't know, like Indian cuisine restaurant nearby me. Yeah, it'll show you the, yeah. But now it, you know, like we are not sure what search queries these chat bots are doing, even if they try to go on Google. 

Joe: Sure. 

Mehmet: Is there, is there a risk of become some business becoming invisible? This is what, you know, why I want to read.

Joe: Um, well, first of all, I do agree. I do agree. Like the, like I said, in the early days, I think there's not that big of a difference. I think these technologies are still maturing. They are statistically very large, right? Lots of users, billions of dollars in, in, in revenue. But the truth is they're not that much farther advanced yet.

That's why the query fan out to me is kind of like the biggest nuance that's, that's changing. Um, but you're right. Like they are in some of them, not [00:11:00] every step along the way, but they are going and just looking at traditional searches for those kind of queries. So right, you get the chat. You get the query fan out, you use Google for those, you know, specific query fan outs, but you're still getting that optionality there and the, the variances, um, I think that's just like any other technology, like they all start somewhere that's not so different.

Um, anything that's too radical usually doesn't get adopted. Um, and, and in addition to it, like there's. Years, 25, almost 30 years of precedence of Google's success. So like, why would you rip it out and try to rebuild it? Right. Um, however, I do think that's going to change. I think that that is actively changing all the time and they are building their own models.

It's just not maybe mature enough for them to, uh, overhaul yet Secondarily. There is still opportunity in those queries that I think is not being created yet. So let me put it this way, those blogs that you might have made about, like I said, like auto repair near [00:12:00] me. So maybe you're, maybe you're writing a blog about, um, you know, why you need to do maintenance before road trip, or why you need to check up on your, uh, BMW's oil more often because of European vehicles and their problems, like things like that.

Are not gonna fall off the face of the earth. But I think there's this new opportunity which we're doing with service stories and we're seeing a lot of success, which is, um. Very nuanced, hyper-personalized content. And that is, uh, for us specifically as a company, we started in auto repair because of my family's business.

We've then expanded into services more broadly, but I'll give you the auto repair example. Mm-hmm. The way I stumbled upon this was that I was trying to work on programmatic SEO for our shop. If you don't know programmatic, SEO, that's just basically you have a, let's just call it an HTML template. You have a H one, H two, all your sections, and like every section header has a variable in it or a couple, and it just switches out for [00:13:00] every different pattern you have.

They're pretty static, ele, you know, pretty static documents that you're just changing slight variables and that's why programmatic is, it's good to get something out really far, you know, and big, uh, volume, but also can be risky if you don't do it well. It looks really. Programmatic and Google doesn't love that.

Um, so anyway, I was starting there. That's what I thought. This is how we're gonna make content for ai. It's, it's a volume thing. And then what I noticed when I was watching day to day is every day my service writers, that's the people at the front desk who help you when you come in about a car problem.

Um. They go out, they look at somebody's car, they ask that person like, what are you smelling? Uh, what are you feeling? What are you hearing? Then my guys go test drive it. They take those notes. They then go make a list of all the work that needs to be done, and we sell it to the customer. Okay? What I realized is that that data is just sitting in our shop management system.

The same as maybe all of your work, your tickets sit in Jira or something like that, [00:14:00] right? All the details of what you need to do and why you're doing it. And I thought, well, every single customer has a different problem. Every time there's a new work comes in, it's a new story, slightly different, even if it's the same vehicle, same year, make and model.

What if I could just turn every ticket into a story? And that's how we created service stories. That's the pattern. But imagine that we're now doing programmatic, but it's, it's not programmed. It's programmatic in the sense we're getting large volume, we're getting different keywords, we're getting all these different conversations.

Literally, I have customer conversations and notes in there, right? So whereas traditional SEO, I'm fully guessing, I'm looking at hfs, I'm looking at emra, trying to guess keywords, looking at trends, all that, and it make content. Maybe I'm looking at Google Search Console, whatever it is, to see what's working for us with our pattern.

We don't have to guess. The customer came in, they said this, they had that problem. And I know from years of working that if they have that problem, a hundred [00:15:00] other people have the same or very similar problem. And so all I'm doing is taking that, reframing it as a question, creating a story, and then that content patterns out into the same type of queries that someone picks up and now voice conversations to chat GPT.

Right. So it's a different kind of keyword setup. It's a different kind of query and. If you go and look up the articles we've created for customers, um, what you'll see is that like it doesn't rank super well. It often doesn't rank for like the big keywords, but what it does hit really well is those very long tail keywords.

And sometimes we will be the only professional result on the page because actually to create content like that historically would be fiscally irresponsible. And so now with our AI software, you can do it at scale for a price. That's a fraction of what you would pay an SEO company to go and do it by hand.

And you are the only one to own that anchor. And I believe that's [00:16:00] the future of SEO is figuring out how to find actual customer conversations. And then match the kind of queries they're gonna be chatting in the chat GPT, the same as we used to try to guess the keywords, the short tail keywords, and match those in our blogs.

Does that make sense? 

Mehmet: A hundred percent. Let me ask you quickly as a follow up on this, which is, you know, it's an eyeopening, I think, for many businesses Yeah. How they should start capturing these important. Conversations that they have today with their customers? 

Joe: Yeah. Great question. And that is the challenge that we're running into is like there are some.

Service businesses who are really well set up for us to plug in that pattern is just automatic. It's part of their daily behaviors. And then there's others where either they don't take notes well enough or more common. They have those notes, but they sit in a different system. Or they sit in three or four different systems and so now we need to combine them [00:17:00] or imagine, you know, go back to the Jira conversation point I had there.

We actually do work with some creative services agencies, uh, software development and um, photography, you know, videography agencies where they do have more of that set up. And that's actually, um. I don't wanna say it's harder for us, but it's, it's more complicated than the original pattern we found, which is just in a shop and on a receipt.

Now I have to pull, not just, uh, you know, JIRA work items. I'm not gonna pull every ticket. I may be, I'm pulling at the macro level, you know, maybe I'm pulling at the project level or something and getting a summary and summarizing that project. Um. I think that's the challenge for us technically and what that means for businesses and what we're seeing is, um, you know, really consider where you store that information, how it's accessible and capturing it.

I think capturing, it's easier than ever before. You know, uh, mm-hmm. Maybe even a few years ago, what we're doing wouldn't really have been possible because people weren't capturing the [00:18:00] front end of it. They weren't capturing the customer conversation very well. Um, we do have a firm that's using it by submitting their sows or statements of work, and those statements of work say, what are the customer want?

What are we doing all that. But like, you know, commonly, where does all that kind of information sit in one system? It just doesn't. Um, but you might have systems that you're now tracking call recording, right? 'cause you're trying to optimize your sales conversations. So now you do have all those, and those conversations are tagged and they're, they're pretty well structured and they could be pulled in.

Or maybe it's sitting in your Salesforce, you have a CRM where you've been talking to customers and you've been entering notes, and then we can pull that in. Um, or, or maybe it's through, you know, granola or different call recorders that you're capturing the transcript and you're turning it in and we're, you know, integrating that with some other data.

Um. It's not all perfect today. I'll tell you that There are some businesses that are more complex than others, but I would say even two or three or five years ago, this type of work wouldn't [00:19:00] have been really able to be captured. 'cause it would've been so time consuming and tedious when it just wouldn't happen.

And then secondarily, uh. There wasn't a need. We didn't have AI search as big or as wide as it is, and so I think we're just hitting a really interesting timing where those two factors combine. Um, but I do think that's gonna be the future search is like really programmatically leveraging actual institutional internal data and customer conversations to match what your content is.

Mehmet: Right. I'm gonna, you know, quickly, uh, kind of, again, I follow up, which I didn't prepare so. Why, you know, still relying on SEO while if I have all this mm-hmm. Why as a business I don't kind of a, you know, um, have my own chat bot, which trained only my on my own data. Yeah. So this will be seen as, you know, like truly, [00:20:00] um, you know, authentic.

And the reason I'm asking you this, Joe, because. You know, when people are discussing what's happening in AI today, the biggest fear and, and you've been in Silicon Valley of course before, and you know, like everyone talks about it. 

Joe: Yeah. 

Mehmet: People are seeing, like, you are more and more concentrating power in platform, which are already large.

You are putting all your eggs in one basket. 

Joe: Yeah. 

Mehmet: Um, and you're relying on like kind of a single way to. To, to, to get noticed and, and all this. Yeah. So, so how, how businesses can do this kind of, um, I would say balance between these two, uh, school of thoughts? 

Joe: Um, maybe let's talk about that for a second.

So are you asking like, businesses committing to Claude? A hundred percent. Yeah. Or to open ai? A hundred percent. And like, yeah. Um, versus what, like versus using a tool like what we've built or what do you mean? 

Mehmet: Yes, yes, yes, yes, exactly. 

Joe: [00:21:00] Um. I, it's always a risk when you concentrate yourself on one platform, but you could say the same about Google.

Or, you know, Microsoft Office, you know, all of those, it's all concentrated. Um, perhaps you could argue that these tools are more concentrated. I don't know if that's arguable in regards to Google. Um, you know, Claude, yeah. You might connect your email. You might connect this and that, these 10 other products, but Google has, excuse me, jeez.

Google has built. Over 500 different products over time and just Google, you know, search Gmail. Maps, YouTube, um, Google Business Profiles, you know, those are some of the bus biggest platforms in the world, and that's five of their 500, you know? So, I don't know. I think it's just new, new competition in [00:22:00] my brain.

Um, and actually some, somewhat, it's not even new competition because you're, you're plugging in your Google, right? So actually Google still has all the data. Um, you're just now connecting it in a different way. And you have like this middle. Middleware that sits and helps you manage it all. So, uh, is it a, a complication?

Is it a problem? Sure. I do think that that's problematic. Um, but in the large language model sense and in the big platform sense, I don't know that there's a lot of optionality. There's not a lot of choice. Um, I do think that will change over time. 

Mehmet: Mm-hmm. 

Joe: I also believe there will be, um. I fundamentally believe there will be a world of SMB softwares that are gonna come out in the future, like, uh, competitors to Google or to others where they don't have a billion users.

They have, you know, 500 or a thousand users and they're a sustainable business and they pay their bills, um, and they're profitable, but they're not out to make a billion dollars. You know, just like a, you know, my family's auto shop came into [00:23:00] automotive. In, uh, 2002, you know, cars have been around for what, 70 years or more, and it's kind of crazy to think about that, right?

It's like they just came into a market that was well saturated, um, plenty of competition, but they stood up and they've been around for 25 years. Um, I think the same can be said for software in the future now with these tools and so. I see the concentration of power today. I have a feeling it will diffuse in the future.

Mm-hmm. Um, large language models are good for a lot of things, but I think you're seeing people now start to build small language models. Mm-hmm. And with the capabilities of ai. I think it's very realistic to manage a, a software that's very niche, maybe even very hyper-local and, uh, sustained by, you know, three to five people and really powerful and useful for its users.

So, I don't know, I, I've, I've thought about that a lot as I've gone through my journey and, and tried to fight and push back on big tech. Um. I don't know that there will ever be a world where they just fall [00:24:00] off the earth. I also don't believe that it would be responsible, whether it's the United States or some other country, to go and regulate them to the point mm-hmm.

Where they can't, they can't operate anymore. Um, they do have a lot of impact on our world. Like imagine, uh, AWS goes down for a day. We, we've seen that in times, right? Like, uh, or imagine Google goes down for a day like. Business just goes off the rails for a day, basically. 

Mehmet: Right. 

Joe: Worldwide, 

Mehmet: right. 

Joe: Um, 

Mehmet: a hundred percent.

Joe: So, so yeah, it's, it's a, it's a coin flip. It's kinda like, well pick your poison, I suppose. Huh? 

Mehmet: Absolutely. Joe, one thing I want to go back to the way how ai, these LLMs, you know, they, they choose, you know, the right thing to, to put in front of us. So, um, so it's like not above the page rank anymore, it's like, hmm.

You know, trying to determine who gets, you know, sitted, right? And uh, yeah. The question that comes here. How, because [00:25:00] we know also about the hallucination. We know about, you know, the, the, you know, all, all the, all the issues with AI that we've seen and still they are there, uh, you know, from time to time. And even we are starting to see now with this agent tech AI in, in the world of, of the broader web and, you know, the way we are trying to give AI 

Joe: Yeah.

Mehmet: Power to do things on our behalf. Mm-hmm. So what would be the different differentiation between. A solid answer to a query versus, you know, and I, I see it more in chat, GPT may, maybe Claude has it, but I never saw it at Claude at the same rate. I saw it at chat GPT when. It tells you like, yeah, this is what sounds right.

And then when you go and challenge it, it goes and tell you, oh, you were right. Let me fix that. So, right. So, so 

Joe: yeah, 

Mehmet: it's about like, you know, the trust, you know. Sure. In the AI tools, what's your take on that? 

Joe: Well, I think just like, uh, bad information. Like [00:26:00] some people are just gonna blindly trust it. I mean, you could argue Google has the same issue.

Uh, yes, a hundred percent. If you don't know what you're looking for, it's easy to just take the first result. Um, I think it's just, you know, I advised the antitrust cases on Facebook and Google, and that was the thing I brought up to them. This was seven years ago. They just thought I was a nutcase because they had no idea this AI search was coming.

Right. But I was telling them, I'm like, I saw this. It's gonna come like, and you're not gonna get 10 million results. You're gonna get one answer. And psychologically people can't remember more than. Six data points. When, when an answer comes, that's like when you get a phone call from an automated service or you get their auto system, it always has like three to six options.

That's psychologically because people just can't remember more when they're on the phone. So. We are going to consistently have like less results, and that's just gonna be part of it. The the thing is, yeah, how are you in that result, if at all. And the other [00:27:00] part of that that you could really consider is that nobody ever went past the first page in Google.

Like, like a statistically insignificant amount of people went past the first page in Google. So like speaking literally in that sense, if everybody, I think it was. More than 70% of people were going to the first three results. Then, I mean, like the user behavior is already there that yes, you have 10 million results, but people are actually looking at the first five or six really anyway.

Right. Logically speaking, whether we like it or not, that's just kind of how it's been happening. So I don't know exactly how that plays out, but what I'm actually interested in, to go back to the question, like how does this change over time? I think when I look five and 10 years in the future. Age Agentic Web today is mostly a bunch of technology people talking to each other and saying, yeah, I'm gonna use a bot to have my work done for me.

Right? Like, my mom and dad aren't using it. A lot of people down the street from me aren't using it. Um, and, and it's not 'cause they can't, it's 'cause they just don't care or they don't know about it yet. It's not aware in the public [00:28:00] yet of, of exactly everything we see day to day. Where I see this going though is Agentic Web is starting literally in a web browser.

Some of us are voice chatting Right now. Google Search just released globally Google Search Live. I don't know if you saw that or not, but now you can take a phone, you can stand it up and let the video watch you do things, and you can ask it while you're working. Like I could be under the sink working on a plumbing issue and say, okay, what is this thing?

And it's pointed up and it's talking to me. Right. We now have a multimodal search and an ai. Mm-hmm. That can almost act like a human assistant, but has no legs. It's still from my phone. Where I see this going over time though, is, um, at some point we will likely have humanoid robots in the house. Uh, we, I already see it like in China.

I see it all over in hotels. There's bots everywhere. Mm-hmm. Robots carrying things or food or whatever it is. It's the early days. The next phase of that too, um, we're, we already know that there are, um, appliances in the home [00:29:00] that are already using, uh. AI to recognize when there's a problem with it. Like my microwave can, not mine literally, but there are microwaves, there are fridges that can now self-diagnose, right?

Mm-hmm. I imagine a world where those things are just connected to the internet, and I have a checkbox where I can say, Hey, when you have a problem, call a technician for me. Right. And so now you have your fridge that's connected to the internet, knows there's a problem. It says, okay, I need to call a doctor for myself basically.

Right. Well, 

Mehmet: right. 

Joe: That's where I think Ag Agentic AI comes in full force now. We have bots talking to bots and it's not just like an AI chat bot from my browser. It's literally like my, my fridge that's calling an appliance repair person. Or it's searching the web and trying to find that. So I think there's gonna be a huge shift.

And this is where also like the sort of stuff we're doing, service stories I think comes in is um. Those, those fridges, the microwaves, the cars, whatever that have problems, I would assume they're not gonna be looking for human style search queries. They're gonna know what the problem is. They're [00:30:00] gonna be looking for someone who's already done that type of work and an example of that work being done really well.

And then they're gonna pick the person based on the actual results. 'cause they're a bot and they're built to optimize. They don't care about the beautiful pictures on your websites or the the video about your company history and all that. They're logic creatures. They want math, they want precision, they want optimization.

They just wanna know, you've already done this job, you've done it 10 different ways to five different types of vehicles and you've been successful. And it looks like you could apply that to me. So I believe you. I'm gonna book a, an appointment and it's all automated, end to end. That's where I see Ag Agentic web going.

Um, and it might be a while out before we really get there. But, um, somewhere between there and now we're gonna start to see all of it being built, and that's where I think the future search really starts to change. 

Mehmet: So we go back to kind of, uh, static, static HTML area, uh, era, you think, Joe? 

Joe: Um, well, I think you're still gonna have websites for humans, right?

[00:31:00] Mm-hmm. I don't think those are gonna go away, and I don't advocate that people get rid of them. But what, what we are doing at service stories is we are helping build AI optimized websites also. 

Mehmet: Mm-hmm. 

Joe: So I imagine a world where you have www for the people and you have, um, dot, ai or MD for the bots. And you just, you just change your robot, TXT and you say, if this, then go to the bots.

If you're human, go to www, you know, kind of thing. And, um, uh, you know, you could also do that through content negotiation if, if you have a custom built website. But the fact of the matter is, like, I. 90 plus percent of the web is, is built on some kind of template, you know, and you have a lot of people who built a website that don't understand what you and I are speaking about right now.

And so you gotta make it easier for, for them. So, uh, I don't see people just like throwing away their website that they invested $30,000 in and, you know, flipping to a new one. Like, I think there's gonna be, um, band-aids basically, you know, to fix problems until we have like a big wave of evolution. Yeah.

Mehmet: Right. If I want to speak a [00:32:00] little bit about the ethical aspect here, and you know, because at your intro you mentioned about, you know, this humane, uh, kind of, uh, use of technology and making sure that it serves us as the best. Yeah. What do you think? Some of the company that might try by themselves to optimize for AI visibility do, um, wrong from maybe the SEO area or maybe, uh, someone told them you do this, but it affect them even maybe on ethical level.

So. Because, you know, there's a lot of of things that, you know, you can throw to ai. We, we, we talked about it a lot, especially, you know, last year and the year before, like junk in, junk out, right? Yeah. So how, how we can avoid this job because there's an ethical aspect and, and to your point. 

Joe: Yeah. 

Mehmet: Going back to an example of the fridge calling a technician.

Yeah. If that technician have put [00:33:00] like raw data about his success with, you know, a lot of people and how, you know, they, they are the best. Yep. They are gonna say, oh, like these guys looks like. Like, you know, good, good ones to, to rely on. 

Joe: Yeah, yeah, yeah. 

Mehmet: Let, lemme bring, let me bring them in. So how, how, how, how we can avoid like these mistakes.

Joe: Sure. Um, well, I think mistakes versus intentional actions is, is different, right? Mm-hmm. Um, we're already seeing people who are just populating, um, hundreds or thousands of pages just because they can now that AI will allow them to. I think that's gonna go away really fast. I think those, those kind of companies are, I think Google's just gonna take 'em off the internet basically.

They're just gonna rank penalize you so bad. Um, you might have got away with it for a few months, but the algorithms are changing fast. There's been a couple, there's one big update December one that I, I'm just now reading about, just re recently happened also. Um. It's a, [00:34:00] it's an initiative. Uh, YouTube, CEO was out in January speaking about ai.

Slop is their biggest focus this year, you know, so I think AI slop is gonna get, um, cut out over time. And how that happens, I think is more the question, a few things that are happening that are causing problems is, you know, poisoning. Of like LLMs, uh mm-hmm. Intentionally sending in factually incorrect information or, um, manipulating maybe competitors because you're sending out content about them that is factually incorrect.

Those things, I think, uh, that's an ethical lens at this point. There's no probably great way to protect against that automatically, you know, through a big scan system. Um, there's also prompt injecting that you're seeing where, um, people are putting in malicious code or. Malicious prompts, it's not even code.

Right? It's like, and that's what kind of makes it more dangerous than anything is you don't even have to know how to program now to make malicious prompts. You just speak plain English and, and [00:35:00] make it do its thing. Um, you know, that's why like, I don't do the whole age agentic, you know, Claude bot, any of that.

I have not probed that. Uh, excuse me. I have not plugged that into any of my services because, um, I just don't. Feel like we're ready yet. I think it's too early. If you have it on an isolated machine in a full setup, Hey, good for you. You're in a good testing environment. But like, ultimately we're at the phase where, um, criminals don't even need to.

Get me to do anything, they can just send me an email. And because I have auto scan on my email, you know, my, my email's gonna read it. It's gonna do the action before I ever open the email. That's crazy to me. Um, and so there's some of that also happening in search, right? Where you're getting malicious prompts, uh, into the search results that are then triggering other things.

Um. That's another one where I think we, I think that will get cleaned up faster because I think those are a little more obvious than, like, if I'm speaking bad about a competitor or something, is kind of hard. The, the big scale volume of stuff is for sure easily to flag, um, just pattern [00:36:00] wise and everything.

Um, but then there's also, yeah, just like coming out and, and creating kind of content that. You want, but you don't actually do. Right. I think, and that was kind of part of your question too, is like how, what if someone just makes stuff up? Right. Um, and I don't know how to express this other than I believe there will be a future, somehow, some way where Google and others is actually built to verify that the work was done.

And that's, that's part of. Why I believe in what we're doing in service stories so much because we're not making up content. And like we've had people come to us and say, oh, well you, you have data about all these industries. I could just plug you in and you could just make this stuff up for me. And I'm, I'm just thi I'm, I'm thinking there.

I'm like, I literally, I could, right? Literally that could be our business that we just help companies that have no content, just propagate everything based on our historical data from other companies. Um, but then in my gut. I just [00:37:00] feel like that's gonna get banned somehow some way. I don't know exactly yet, but um, I definitely could see a future where there's more like, quote unquote verifiable content that's required.

And if you have your ticket, your work actually turned into content that will exist, that will be forever strong. That will be a huge moat in SEO versus if you just. Propagate a hundred or a thousand articles about work you could have done at some point given the opportunity, but you never actually did yourself.

Um, and so I think the combination of that will be yes, like some kind of verifiable content. Secondarily, um, Google. Chat, GPT, Claude, they're not dumb andro, they're not, they're not dumb. So they're also triangulating by, you know, looking at reviews, looking at press, looking at different third parties to verify that this is a real business versus you just standing up a website and putting it all up and faking it.

Um, I do think that's gonna happen. I know it's happening, um, but I think it will [00:38:00] get. Controlled faster than previous eras because they are aware that that causes problems for their users. Number one, but even more important to them as a business is it causes problems to their ai. It poisons their data, it makes their services worse.

So they are focused on cleaning it up. 

Mehmet: Yeah. So it's, it's their benefit to, to, to, to be sane and 

Joe: yeah. 

Mehmet: Put the right stuff over there. Now, if we, we just like talked about how, you know, websites might have like two interfaces in the future. One for the humans, one for, for the ai. Now what companies should.

Start doing today. Right. So in order, like they can still exist around, I would say in, in three to five years time. Yeah. So what would be the main difference that companies are doing so they can. Really stay relevant to their customers versus [00:39:00] disappearing again with all these ai, you know, whether it's like the agent part of it or you know, the AI optimization part of it.

Sure. What will be the what, what you're advising companies now, Joe, to do? 

Joe: Yeah. I mean. A few quick ones that you can take with you and do today. You know, if you're not already using some kind of schema markup on your website, start to do that. Prepare it for that. I think that that is more important than ever.

Um, I, I put it this way, I actually wrote a Forbes piece about it, um, and it's part of the work we're doing, like I said, so I'm analyzing the performance of it a lot. Um, if your website is not properly structured. I can tell you from looking at a lot of websites that the average homepage costs chat, GPT, Google, whoever's crawling it, anywhere from 60,000 to 200,000 tokens.

Right, because we have a bunch of JavaScript load. We have a bunch of navigation, the bots don't need videos, all that kind of stuff, right? [00:40:00] If you built an optimized version of it, which we're working on, you're looking at like 2000 to 10,000 tokens, okay? Now, this is early days, so we don't have all the numbers to validate the performance yet, but imagine if you were buying a, a part for your car, okay?

And you had an option that you were looking at on the catalog and one option was the part you need and it cost you a hundred dollars. The next to it was another part. It was the exact same part, maybe even a little bit higher quality, and it only cost you $10. What one are you gonna buy? You know? What are you gonna buy as a business owner?

You're gonna buy the cheaper part. Right? Right. This is actually, this is the feeling I had in China. This is a side note. When I went to China, everything was 60 to 90% cheaper than the United States, and a lot of it was higher quality. I was like, oh, this is what happens when you buy it, where it's made. But anyway, um, I think that's, that's a big thing too, that people haven't considered yet and hasn't really [00:41:00] taken, um, hasn't gotten traction yet because, no, it's, it's very early, but I think that's what's gonna happen.

Like I said, optimize for ai. Same as when we had to optimize for mobile. What does it mean to optimize for ai? I think making your site ready and readable. Um, making it so that you have all the details of your work, not just like big keyword phrases, um, making sure that you are staying active, right?

Recency, imagine you're having a conversation with people. We have all done business outside of the city we live in, and how hard is it to keep in touch with those people? It's even hard with the people that are in our city to keep in touch. Well, imagine your job is now to have a, a consistent conversation with Google.

So how do you keep in touch Recency's important? Recency means more blog posts. That means more reviews on your Google business page. That means social media posts because Facebook and Instagram are now indexable. All those kind of things that you need to keep updates. So like three points I would leave with is, um.

Make sure you have [00:42:00] markup for your website. It's maybe not as important today as it will be in the future, but it is important. And those who do are seeing results. Um, make sure that you have the type of content that AI is looking for, which is different, although it's new. Uh, it is different. And make sure that you have a persistent volume, even if you're just doing one a week, like that's better than a lot of people.

But, um. I think more volume to an extent is better. You don't wanna just go nuts and throw everything at it, because then you'll probably get penalized. But like, um, I think if you can get two or three a week and maybe you can grow over time, you know, especially in high volume businesses, you could do like daily kind of stuff.

Um, I do think that will get rewarded at the end of the day when it comes to ai. So 

Mehmet: great. Um, I, I'm introducing something new to the podcast. I never done this before, which is the rapid fire in just three questions, uh, Joe, like quick answers. So, and it's kind of a, it gonna be a wrapping for, for the episode.

Okay. [00:43:00] And, you know, like, this is what I, I was able to, to put in my head. So, SEO debt or Evolving? 

Joe: Evolving. 

Mehmet: Okay. Um, hmm. And trying to remember what we talked about. Yeah, so content versus data, which one wins? 

Joe: Gosh. Um, I think data-driven content wins. 

Mehmet: I like that. And what's worse? Uh, not being visible or, um, misinformation.

Joe: Oh, misinformation for sure. You can always fix the. Invisibility problem. But if you feed misinformation and you get banned, you could be out for months or years. Not good. Uh, 

Mehmet: yeah. I, I think this was a good try. I cannot, yeah, I gotta do it again with my, my, my guest moving forward. Um, finally question. I asked to all my, my guest Joe.

So kind, the CTA call to action where people can get in [00:44:00] touch and if they want to learn more about server stories or maybe get in touch with you. 

Joe: Sure directly contacting me is best on LinkedIn. Uh, that's where I'm most active. Uh, and if you wanted to get a demo of service stories or learn more, you can go to service stories.com and click book a demo in top right corner.

And I would be happy to, to talk to you about it. I'm very excited about it. 

Mehmet: Great. I'll make sure like the, you know, your LinkedIn profile, uh, link is in the show notes with the website of server stories also as well. Jo, I can't thank you enough for this really, you know, eye-opening, um, discussion today. Uh, and that's why I love about doing the show 'cause I myself learn new things and, you know, the idea of getting.

Your own business data put into stories, and this is why you call the, you know, and this like, create all these, um, you know, um, content from, from, from it and making it something authentic for people to come and find you. This is was an [00:45:00] eyeopening for me, uh, and thank you for sharing also all your insights about, you know, what's happening around the web.

The AgTech world and you know, um, how the world of LLMs and AI is evolving, and this is how I end my opposites. This is for the audience. If you just discovered this podcast by luck, thank you for passing by. We're trying to grow as much as possible, so please. Share with your friends and colleagues and also subscribe, and if you are one of the people who keeps coming again and again, thank you very much for the support, for loyalty.

You without you. I repeat this at the end of each episode. I know this, but I need to thank everyone who's referring other people to the podcast. And because of this, since last year, since 2025, every week we keep changing countries. So every week I see the podcast is ranking in the. Apple top 200 podcast chart in one of the countries around the world.

And, and the good news is we keep changing. So we keep visiting new countries every week. So at the time of this [00:46:00] recording, 31st of March, where in Slovenia? We're in Tanzania and we were in Cayman Island. So maybe who knows next week where we'll be. So thank you very much for tuning in and see you very soon in your episode.