The Future of SEO in the Age of AI Search

0:00You know, over the years, you've heard me talk a lot about marketing systems.

0:30Tape Marketing 3.0. As a dad, I couldn't be prouder, but I want you to check it out at dtm.world slash unchained and get ready to take your strategy to a whole new level.

0:53Hello and welcome to another episode of the Duct Tape Marketing Podcast. This is John Jantz and my guest today is Todd Sawicki. He is the founder and CEO of Gumshoe AI, an innovative platform at the forefront of AI-driven search and discovery solutions. With a background in digital media, marketing technology, and leading high-growth startups, Todd is known for his deep insight into changing landscape of search. We're going to talk about SEO. We're going to talk about AIO, AEO, all the other O's. As long as you don't call it GEO, what you can tell the person who came up with that had no background in marketing because I'm sorry, the minute I've been in the paid landscape, the minute you see the letters GEO, you instantly think of geo targeting like, hello, people.

1:39The last thing we want to do is make anything more confusing than it might otherwise be. So and so today with that, Todd's on the show. Welcome, Todd. So let's I mean, I kind of laid that out a little bit. You've created a tool that is really taking advantage of some of the changes that are going on in marketing today, especially around search. So maybe give a high level kind of in your view. Let's start with the basics. All this stuff we're hearing about, geo for one, AO, AIO, all those kind of things. I mean, what does it all really boil down to for the typical marketer or typical business? It is a good question.

2:19So I think we all woke up a year ago and with the rise of zero click searches with AI mode in Google search taking off and we began to see Google traffic starting to decline. And at the same time, if anyone was sort of looking at their GA4 analytics or whatever they're using, they started to see, oh, look, I'm getting this new basket of traffic from ChatTBT and others. And so AI and sort of looking at that. And so the AI search is taking off. And so as a marketer, suddenly you had to start paying attention to this new thing called AI search. And so fundamentally, we look at it as, you know, marketers want to understand what the hell are LMs saying about me?

2:59And then from a product standpoint, we like to say, yes, we help marketers understand what LMs think about them and their brands and ultimately what to do about it. And I think that's one of the interesting things is there's a lot more you can do about it because AI search is a fundamentally different platform and approach than traditional search. And really, in many ways, I think AI search is solving a lot of the problems and complaints as end users we've had about traditional search. And then there's downstream implications for marketers and how to think about how you work with those platforms as a result.

3:30Well, and I think you're hitting on one of the things that I try to get people to understand. Everybody always goes, oh, we've got these new platforms. But what they fail sometimes to recognize is that the buyer behavior is changing because of these platforms and how people, what their expectations are, how they now go to even to Google. I mean, I'm seeing people do this. We used to put in these nice little compact searches. Well, I'm seeing people put in these very long searches now. Exactly. Very high intent, you know, very filtered almost because they know they can get AI overviews and things.

3:59And I think that change is really what we really need to adjust to, right? It's not necessarily the technology, is it? I agree. Users have fundamentally changed. And you probably hear this even anecdotally amongst your friend sets. Like you start kind of experimenting with chat TPT or perplexity or whatever it is. And you're like, you ask it a real deep question that you know is very frustrating to get answered in traditional search. and you would have to click through 10 things and it was just a pain in the ass and took a little time. And where now you get a pretty good answer most of the time right away and it fundamentally changes the experience.

4:30I mean, we're seeing dramatic changes, especially in complex areas like B2B type searches. It's a great use case when you're researching very technical things. You're researching like more long tail areas for traditional search work wonderfully in the world of AI. And I think the other thing that traditional search really did a poor job of, and it really shows up in AI search, is AI search does a phenomenal job of personalizing its answers for you. And that is one of the things that, even in terms of our own product and platform, the implications of that are very interesting. And so as an end user, would you imagine, think of the LLM as you walk into a shoe store, and there's a wall of 500 pairs of shoes behind that salesperson as you walk in.

5:17and the LLM is the salesperson. And so you're trying to, what's the right pair of shoes? Well, Google, you do it. It doesn't really ever answer. I need a new pair of shoes. You would never, like Google just would struggle with that. Or give you the most popular shoes or whatever. Or give you the most popular one, exactly. Just give you the most popular one. But the LLMs are really trying to understand, are you a runner? Are you a hiker? And if you have an account, you register, they're building profiles of you, interestingly enough, right? The minute you put your email in, it knows where you work.

5:40It knows what you're affiliated with. And so as a result, your users are seeing that they're really, there's a value for that relationship between you and the LMs. It learns more about who you are. It discovers things. It's trying to personalize the answers. And so it therefore can give you a better answer and really help you in a way that traditional search never quite got to. You know, and one of the things that I get business owners pretty excited about, because a lot of them are going, oh, is all hype or like, do I got to really do this? Or am I really going to get AI traffic or not get AI traffic?

6:12So all these questions and all I do show them analytics. And I am able to demonstrate that to them, the people who come from AI stay on your site 10 times longer and convert seven times more than your paid ads, more than your organic traffic. And a lot of that, I think, is just what you talked about, because they are doing the filtering themselves. And if they get to your website, it's because you had what they wanted. Right. No. And we're seeing stats on the B2C. We typically see a little bit less than 7x, probably more in the range of kind of two to five X increased conversions. On the B side, we're seeing increased conversion rates up to like 20 X better.

6:51Because again, they're down the funnel. Because the right way to think about, if you think of from a marketer standpoint, let's think about the classic marketing funnel. There's discovery, then consideration, then conversion. Google managed discovery and then handed you off to websites to manage consideration, like your own website, some third party writer, whatever it might be. but AI is trying to do not just discovery, but manage through the Q and a process consideration as well. And then hand that user off for conversion. And so that's why you see these higher conversion rates. They're further down the funnel.

7:21AI has managed that. Now from a marketing standpoint, you're now your challenges. Oh, I need to manage AI differently because now suddenly it's the one selling my product. And I think that's the fundamental shift here as a marketer is you have to going back to that, that shoe store analogy, that LMM as a salesperson means you're going to have to manage. that person, right? That's now your job. Whereas SEO, and I think this is one of the other big changes, SEO is a very technical thing, like link building. And remember just the ridiculous debate we had for years about, is it a subdomain or a folder, right?

7:53Is that marketing? No, that's a very technical thing. And any non-technical marketer, whenever that discussion, and by the way, for those who don't pay attention, that went on for years. It was like a reds versus blues sort of battle in the online marketing sphere. But a very technical thing, not marketing based at all. And I think the difference is for LLMs, it's much more of a, oh, how do I teach the LLMs what to say about my product? Just like I teach, you know, a salesperson at the front of Dick's Sporting Goods store, kind of the same way. And so it's now it's much more of a product marketing exercise than it ever was with traditional search.

8:26And I think that's the other thing is you're going to have to think about how you talk to the LLMs and how you market to them. Well, and this gets at the crux of, you know, a good salesperson is trained on, you know, all the objections of all the questions they're going to get right and so now all of a sudden our content has to be answered correct you absolutely so one of the things so gumshoe as a platform has been we publicly launched it about six months ago and we've already worked more than 3500 marketers have signed up we've already generated millions of prompts on behalf of marketers so they understand what lms say in response to these prompts and as a result we're able to analyze those response i I think it's like 10 million answers that we've analyzed.

9:08And then you really, you start to see patterns in what they're doing, but they absolutely want you as marketers to provide them kind of sample question answers back. Like if you, one of the fascinating things about LLMs is they actually link. They prefer the number one source that they link to for product information are brand websites. And then within that, they link to product description pages or PDPs or product detail pages, whatever description you want to use. PDPs, FAQs, knowledge-based articles, how-to sections. They love that sort of informative how-to answer questions for them. And they use that as a guide.

9:46Now they process their own way. They kind of regurgitate it in their own way, but they want to use that as a basis. So you're right. You have to, just like you train that salesperson on rude Q &A, you're doing the same thing now with the models. Which I think is interesting to marketers when they start seeing that understanding, like it's now a marketing exercise and not a weird technical link building subdomain folder esoteric discussion anymore. Yeah. And one of the things that we have seen, because, you know, I've always believed that you do content right, you're going to get rewarded by the search engine.

10:14Yes. We've been doing content right, in my view, you know, hub pages, structured content, FAQs, table of contents, summaries, schema. You know, we've been doing all that stuff because it was good content marketing. Well, the LLMs and AI are actually rewarding us for that work right now because we ranked high in Google. We are now ranking higher in AI overviews and in ChatGPT. Are you seeing that as well? So if you don't have content online, it is hard for the for AI to even know you exist. Well, yeah. And so that's sort of step on. You'd be surprised at the lack of content out there. No, I would not be surprised.

10:53All right, well, you sell it. You sell these programs. But I think it's because everyone probably thinks everyone's done content marketing. That's not always the case. I always love it when we go to work with a new client and they say, yeah, well, our SEO firm is doing this for us. And it's like, what are they SEOing? There you go, exactly. There's no content, there's nothing else. And so the difference is here, you mentioned like you generated content. The difference here though is there's a subtle benefit and you kind of address this. I'm gonna call it what you said, which is you're getting rewarded.

11:24But what's interesting is Google, it was rewarding popularity, not necessarily the best content and the most authoritative content. And what LLMs are doing is doing a much better job of rewarding the correct content. So it's, and we have a good stat around this, which is, so we look up the traditional Google rank of all the URLs that are cited by AI and its answer in its justification for its answers. The traditional Google rank is below 21, 50 to 90 % of the time, meaning page three and beyond. So it's pulling out these. So it is looking at some of those that traditionally link to content SEO, but it was always these deep links.

11:58And the problem with traditional searchers is, you know, we kind of generically use the stat. One out of a hundred people go to page two on Google. One out of a thousand go to page three. One out of 10 ,000 go to page four and no one goes to page five. That's where you hide the dead bodies, right? That's right, exactly. You hide the dead bodies. But AI, to my stat, 50 to 90 % of the links they surface are in that sort of buried in, because they have AI or machines, they have infinite patience. So what they're good at doing is finding authoritatively correct, like we like to say canonical information.

12:29And so as a brand, all that work that maybe struggled to get surfaced in Google because it just wasn't as popular or you're losing out to people buying links. Now they're really, to your point, really rewarding good content. Good, highly valued structured content. And so it's sort of the, it's paying off 10 years of work finally. And so the people who maybe struggled to get some of that popularity in Google, It is absolutely paying off in AI overview, AI search and AI overuse and things like that in a way that you always prayed and hoped for as a content marketer. Like your day has come. Producing great content is a payoff and it's happening.

13:01And I think that's really fascinating here, which is people are like, oh, with the rise of AI slop. No, the models want good content and they're good at deducing what is good content. AI slop will not get ranked and you have to they want authoritative information. And so that's content that will get ranked in AI search and then drive traffic today and tomorrow agentic purchases. Right. You're ultimately trying to drive some of the conversion more and more that AI will be driving that itself. Like Perplexity's browser will load a cart for you today. Right. It's loading products that's picking on your behalf into that.

13:31So that that future is coming fast and furiously. And so I think that changes is sort of fascinating to see when you look at what's happening. Now, the other stat about what's really fascinating here is, OK, what if I don't have been produced 10 years of content? Am I screwed? Well, one of the other facts that we've seen is that the average age of a sighted piece of content is only 86 days old in AI search. And that's falling 10 to 15 percent quarter per quarter. Now, there's a caveat there, which is it doesn't have to be originally published. It just has to be updated. Like the AI will look at content that's older, but as long as it's been updated and you note that updated date, it will value that as well.

14:13And so, and that 86 days is falling 10 to 15 % every quarter. So today it's 86 days. Next quarter it's gonna be 78, 70 the quarter after that. And so you get faster and faster. So you're gonna have to be doing a lot more work around content, maintaining it, updating it. It's not a publish once and walk away model anymore. It's like being a constant refresh. And so, but the good side of that is you're just starting out. We've definitely seen this with people where you can impact the results well within a 90-day window where traditional search, that was almost impossible. And so there's definitely a, don't wait, get started.

14:45Yeah, well, but I was also going to say another best practice for years has been repurpose your content. And so, I mean, now it's like repurpose your content in a specific way, you know, add FAQs, you know, to that content, right? But I think that's what you're saying is should be very helpful for those people that just kind of wrote the 100 one-off blog posts. It's like, no, now go back and make that pay. You know, over the years, you've heard me talk a lot about marketing systems. Today, I get to share something really special. My daughter, Sarah Nay, who also happens to be the CEO of Duct Tape Marketing, just released her first book.

15:20It's called Unchained, Breaking Free from Broken Marketing Models and How Small Business Can Finally Take Control of Their Marketing, Lead with Strategy, and Scale with AI. I know, a mouthful, right? But it's everything we've learned taken to the next level. In fact, we're even calling it Duct Tape Marketing 3.0. As a dad, I couldn't be prouder. But I want you to check it out at dtm.world slash unchained and get ready to take your strategy to a whole new level. Let's talk specifically about gumshoe. I know that's what you want to talk about. But first off, I have an account. I've played with it and it is seemingly incredibly complex what you've built.

16:04And so my first question is, my first question is, where did that come from? Are you a mad scientist or did you hire people or how did you develop that? So we have a team, right? We have a team. I've been in digital marketing tech for 20 years in my career and got involved in, and really the common theme has been around customer acquisition as it turns out. And I even view the purpose, we only care about AI search as marketers ultimately because it can drive business. It'll drive traffic and revenue. So fundamentally, it's a... And so I, 20 years ago, got involved in toolbars and search. Then I got into the social marketing landscape, just as that was taking off like 2007 to 2012, and then got into paid and built a DSP.

16:43So in the programmatic space, and then was playing e-commerce and Shopify's ecosystem, building customer acquisition apps in there, and then ultimately transitioned here. And it was sort of the space of a year ago was talking to marketers. And again, the beginning of this conversation around AI search and the rise of that. And if you're a marketer and suddenly the channel you're relying on, Google search, falls off a click. And for some key keywords, I heard 30, 60, even 90 % declines in traffic, even on the paid side. Like it just, Google is sacrificing even paid traffic on some keywords. So that's an existential change in the landscape.

17:18And then as we started thinking about this in terms of working with marketers, you're like, well, you know, to what I said earlier, Gumshoe helps brands understand what elements think about them. and then what to do about it. Well, where does that come from? Well, if you're a marketer, you can't just log into ChatTPT and find out what it's saying to you because I don't know if anyone should go watch the season premiere this fall's episode from Boulder Natives, you know, the creators of South Park. The first episode this year, one of the main characters' dads is like falling in love with ChatTPT because all it does is flatter him.

17:52Right. And it says like every idea he has is wonderful and it's great and he's got some, he's trying to start a new business and his wife gets all pissed off because he's constantly going to ask chat tbt and says see i'm right you know his wife's name is sharon see i'm white sharon chat tbt says i'm right and he's like no it just says it to everybody and so as a marketer you can't just log in and ask chat tbt what it thinks about your business because it's going to kind of lie it's going to flatter you it's going to say the most optimal thing it can because it by the way the minute you put your email in it looks you up on linkedin it knows where you work It knows your products.

18:22It knows how to answer things. And so then you realize as a marketer, I don't care what LM say to me. I care what it says to my target customers. And so the way that we built our product was around how do you help marketers understand what it's really saying to its customers? And so our point of view is, well, how do we get in the shoes of that customer? And so what we do is we build these personas, which become synthetic users. And right. So those are what are asking prompts of the models. we have a better understanding of how they, how will they talk to, how the models speak to these different customers and those insights of like, okay, here's how, and by the way, the variety of answers between one type of persona and another is fascinating.

19:01And they're absolutely customizing their answers. Like John, you've seen this, right? Just one customer will say, like, just imagine you're a hiker. You're going to get a different answer for the pair of shoes than if you're a marathon runner. And so that makes rational sense as a marketer, kind of understanding this nuance and how it's treating different types of end users using AI search is sort of a fascinating insight. And it's cool just to look at the answers and see what they say to different things. So that's my point about marketers and the messaging and seeing how it talks to different people.

19:27One of my first observations that kind of blew me away, frankly, was I just put in a company's URL, I think is all I did. Correct, that's what you start with. You start with the URL, correct. And it came up with, I want to say eight, maybe it was a little more than that, personas. And they were we had already done that work, but they were very spot on, maybe even a little better descriptions. And what I found was interesting was it actually. All the analytics and search was great, but we actually got some messaging ideas just from that part of it. And that wasn't even the intent. Well, and that's what I mean about it.

20:06I was talking with a head of product marketing earlier today, and I'm like, this is product marketing's moment because AI search is fundamentally a product marketing exercise. And it's a positioning exercise. And when you read those prompts and answers, we hear that all the time because what we help you on, we ask questions and basically ask questions around product areas for your business. And those will give you a set of responses like, oh, we recommend these three companies or these eight companies or these five. and then you see the rationale for those recommendations. And that's great marketing, right?

20:38Feedback. It's what's our positioning? What's our competitive positioning? You show this to any product marketing, like, oh my God, this is like my competitive messaging framework, which you, by the way, what you described, John, A, it's self-serve. You can do this yourself. Anyone can enter a URL of a company to get this. And in like 10 to 15 minutes, you're walking away with a really cool understanding of your product's position in the marketplace, at least the marketplace of AI search, which is meant to be a broader perspective of the world, obviously. But it's, no, I hear this all the time.

21:03It's fascinating. Like it is a total rabbit hole for anyone who cares about competitive or comparative messaging. So the other observation is that, you know, a lot of people that are talking about losing search traffic, it's for, let's say I'm a remodeling contractor. It's they're losing traffic for trends in kitchens, right? Which was not somebody that was going to buy anything, right? They're losing a lot of that traffic because they'd written a great trend article for 2025, right? But that was not going to ever convert. But what's interesting from what you're unearthing is you're unearthing all these really high intent searches.

21:37I mean, the search string is such that it's like, yeah, that person's looking to remodel their kitchen. And I think that's what marketers need to really focus on is that forget about the deals. We do still have to do a lot of things to create awareness, but what we really need to focus on is high intent right now and capturing that search. That is absolutely, I think, a change, which is you're going to go a little bit more down funnel. And because I think you can with AI search. The problem with Google is all those searches were so high level and so generic. It was hard to, you're right, the lack of long detailed searches in Google meant it was hard as a marketer.

22:12You couldn't really target that sort of bottom of funnel activity. But AI is kind of all about that. And even if you ask a generic question, AI will follow it up with a more specifically. They want to know which direction they need to go. There's a back and forth that never existed in Google search that absolutely exists in AI. And anyone who's experienced this when you go to the models, it'll ask for follow-ups. It'll clarify things. It'll make sure it understands what you're talking about so that its goal is to give you the very best answer possible. Yeah, wouldn't that be great? You go to Google and say, no, that answer was wrong.