The Future of SEO in the Age of AI Search  ## Metadata - Author: [[The Duct Tape Marketing Podcast]] - Full Title: The Future of SEO in the Age of AI Search - Category: #podcasts - URL: https://share.snipd.com/episode/ac6e1e91-e998-4f5b-b89e-a3359fc6dc29 ## Document Notes ## Highlights - **AI Search Is A New Marketing Platform** - AI search is a fundamentally different platform than traditional search and changes how marketers must think about discovery. - Gumshoe helps marketers understand what large language models (LLMs) say about their brands and what to do next. Transcript: Todd Sawicki Is a good question. So 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. Anyone was sort of looking at their like 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. So the AI search is taking off. And so as a marketer, suddenly you had to start paying to this attention, 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. And 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. ([Time 0:02:18](https://share.snipd.com/snip/6cc30733-06ef-48b7-8394-128eb841c9e6)) - **Users Expect Personalized AI Conversations** - Users now expect personalized, conversational answers and they enter longer, high-intent queries into AI systems. - LLMs act like a salesperson that tailors recommendations based on user profiles and account data. Transcript: Todd Sawicki 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 lot of time. And where now you get a pretty good answer most of the time right away. And it fundamentally changes the experience. I 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, right, 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. And 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. John Jantsch Or give you the most popular shoes or whatever. Or give you the most popular one. Todd Sawicki 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 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, it 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. ([Time 0:04:07](https://share.snipd.com/snip/7e630ad9-3def-4dd8-8d06-cd11563495cb)) - **AI Traffic Converts Far Better** - Traffic from AI-driven discovery tends to be further down the funnel and converts much better than generic search traffic. - Marketers must now manage AI as a seller that handles discovery and consideration before handing off conversion. Transcript: John Jantsch 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, you know, do I got to really do this or am I really Going to get AI traffic or not get AI traffic? 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 they're doing the filtering themselves. And if they get to your website, it's because you had what they wanted. Todd Sawicki Right. Right. No. And we're seeing stats on the B to C. We typically see a little bit less than 7x, probably more in the range of kind of two to 5x increased conversions. On the B to B side, we're seeing increased conversion rates up to like 20x better. Because again, they're down the funnel because the right way to think about 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&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. ([Time 0:06:00](https://share.snipd.com/snip/e34f29c6-0c68-4cfc-b110-7dc66dd453cb)) - **Train LLMs Like Sales Reps** - Treat LLMs like salespeople and proactively teach them what to say about your product. - Provide structured PDPs, FAQs, and how-to articles so models cite and use your content as authoritative sources. Transcript: Todd Sawicki And remember just the ridiculous debate we had for years about, is it a subdomain or a folder, right? Is 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. Like it was like a reds versus blues sort of battle in the online marketing sphere. And, 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'd 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. And 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. John Jantsch 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. Todd Sawicki 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 3,500 marketers have signed up. We've already generated millions of prompts on behalf of marketers so they understand what LLM say in response to these prompts. And as a result, we're able to analyze those response. I think it's like 10 million answers that we've analyzed. And 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. The 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. Now 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 root Q&A, you're doing the same thing now with the models. ([Time 0:07:46](https://share.snipd.com/snip/9447d59c-14e5-4416-bb0c-df0b60e8ef4c)) - **Quality Content Wins Over Popularity** - LLMs reward correct, authoritative content more than popularity, surfacing deep links that traditional Google might bury. - Good structured content and updated pages are getting renewed payoff in AI overviews and chat-driven search. Transcript: John Jantsch 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. Yes. 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 chat GPT. Todd Sawicki 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. John Jantsch All right. Todd Sawicki Well, you sell it. You sell these programs when I think it's because everyone's think that everyone probably thinks they've all everyone's done content marketing. That's not always the case. John Jantsch 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? Todd Sawicki 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, 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. But 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's, it is looking at some of those that traditionally linked to content SEO, but it was always these deep links. And 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. John Jantsch That's where you hide the dead bodies, right? Todd Sawicki That's right. Exactly. You hide the dead bodies. But AI, to my stat, 59% of the links they surface are in that sort of buried 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. And 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, now they're really, to your point, really rewarding good content. Yeah. Good, highly valued, structured content. And so it's sort of the, it's paying off 10 years of work finally. ([Time 0:10:06](https://share.snipd.com/snip/41ecab09-c36c-480f-b4a3-079f5996f87f)) - **Refresh Content Regularly** - Update and refresh content frequently because AI favors recent or freshly updated material. - Aim to impact AI results within a 90-day window by maintaining and repurposing existing content with FAQs and clear structure. Transcript: Todd Sawicki 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. And so, and that 86 days is falling 10 to 15% every quarter. So today it's 86 days. Next quarter is going to be 78, 70 the quarter after that. And so you get faster and faster. So you're going to 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. ([Time 0:13:41](https://share.snipd.com/snip/9d5ccd79-afe2-4352-a5c5-6bd750dc2e06)) - **Test AI From Your Customers' Shoes** - Don’t measure LLMs by what they tell you personally; simulate your target customers to see what the models say to them. - Build personas (synthetic users) and run targeted prompts to reveal how AI ranks and recommends your product. Transcript: John Jantsch 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. And so 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? Todd Sawicki 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, right? It'll drive traffic and revenue, right? 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. So in the programmatic space, and then was playing e-commerce and Shopify's ecosystem, you know, 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 cliff. And for some 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. And 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. ([Time 0:15:52](https://share.snipd.com/snip/6bca4f33-b6d2-4858-8de4-3a83818ea4d4)) - **URL Input Produced Useful Personas** - John entered his company URL and Gumshoe generated multiple spot-on personas without manual input. - That process also produced unexpected messaging ideas useful for product marketing. Transcript: John Jantsch 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. Todd Sawicki Well, and that's what I mean about it. I 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? Feedback. 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. ([Time 0:19:30](https://share.snipd.com/snip/a774b4a3-349a-454b-b9da-dc488b9b5a65)) - **AI Brings More High-Intent Queries** - AI search surfaces more down-funnel, high-intent queries, enabling marketers to target purchase-ready users. - The conversational follow-ups in AI let brands clarify intent and capture stronger leads than generic trends content. Transcript: John Jantsch 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. 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. I 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. Todd Sawicki 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. You're right. The lack of long, detailed searches in Google meant it was hard as a marketer. You 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. ([Time 0:21:10](https://share.snipd.com/snip/95ab7302-def7-4c77-b3a7-045607a16c14)) - **Use Gumshoe To Audit AI Perception** - Run a Gumshoe report by entering your URL to get persona-driven prompts and real-time LLM conversations. - Use the analysis to identify topics where you win or lose and then produce targeted content to instruct models. Transcript: John Jantsch How does somebody try it out? Todd Sawicki So any marketer, it's publicly available and you can try it out for free. It is you can generate a report about your company and you go to it. As John said, you're going to enter your company's URL. And then from there, what we're going to do is, again, show you what LMS think about your business and product. You're going to select a product that'll generate personas and then we'll generate the prompts that represent the activity that users are having with AI and then run a series of real-time Conversations. We turn those personas effectively into synthetic users. That's kind of the buzzy word. Synthetic is the ultimate. Now, AI buzzwords, it's a simulated user. It's a synthetic user. It's better than that user, isn't it? Yeah, exactly. It's better than that. We'll have a series of conversations with the LLMs. We kind of create those and then we analyze the chat activity and kind of package that up in a way so you can help identify areas, topics of these types of prompts where you're doing well Or you're doing poorly. And then the next step is we also allow you to sort of then generate the content based upon, you know, where your strengths and weaknesses are that through our platform that you can then Host on your site. And the way to think of it is your personas are your predicted customers, who the LMs think are your top customers. And then they want instructions. The content you generate is intended to be or write on your own is intended to be the instruction set back to the models. OK, for these customers, here's the features and benefits that we believe appeal to them and why they want to pick our products. And ultimately, it's going to send traffic back to your site. And then you can help analyze that to understand was it good traffic, bad traffic, what have you. And so the goal of our point of view is to say, again, how do we help you understand what LLMs think about you and then what to do about it, right? Ultimately, how do you capture as much revenue or as much referral traffic as possible from the LLMs? And so that's the way Gumshoe works. You can go to gumshoe.ai. As I said, you just start with the URL. ([Time 0:23:28](https://share.snipd.com/snip/6f609353-1812-401d-85bf-a2b5f2fb621b)) - **Get Competitive Messaging From AI Chats** - Gumshoe surfaces competitors and the reasons models mention them, giving product-marketing grade competitive intelligence. - Those insights reveal positioning gaps and the citations or sources AI used to justify recommendations. Transcript: John Jantsch Either, is I thought it does a really good job at identifying competitors as well. Correct. Because what we'll do is in those answers, we're going to get multiple companies, products, recommendations, and we surface that to know your competitive. Todd Sawicki Great point, John. You know your competitive standing. Are competitors doing better or worse than you in AI? And that's obviously often a key indicator. And then we'll help you analyze where they did better versus you. So you know what's your point about messaging, right? And the product messaging, like what features of a competitor are winning versus ours? Is their positioning better? Is it something else? And that's sort of a great insight is who are all the other companies getting mentioned alongside you? And then we'll help you identify also what were the reasons, like what led to the models answering the way they did? Like what citations and sources. So if you want to do outreach from a PR standpoint, we help you identify the places you should be going and talking to, or even Reddit or Quora threads you should be posting on. We now have a feature where we'll give you a draft post for Reddit and Quora. Again, but it's based upon, you know, strengths and weaknesses that we identified and said, oh, here's the things you should be talking about more to help you get more visibility to AI. And so that's sort of the goal here is how do we help you talk back to AI? So you're feeding it the features and benefits of your products. So they'll talk about your products next time instead of someone else's. ([Time 0:25:50](https://share.snipd.com/snip/bc04d43c-3f58-4c7e-815c-6765946cdcf2))