Topic-Based GEO: The Content Strategy Framework for AI Search

Chat GPT or any LLM has so much personalized context about the user like what type of business it is, what's their goal, what's their preference, what they've used before, what they've already tried, what they failed. What I'm saying is even if you are tracking the exact prompt that a user happens to type in, even that, by the way, is low probability because a user is going to have a conversation. Most people don't treat this like Google and type find me a content agency. Enter. They have a long conversation about their business problems and then eventually that leads to content agency. Even if you get lucky and you type in the exact you're tracking in your tracking tool, the exact prompt that the user is typing in, you're still not going to get the same answer as them because you lack their context. — The companies that are being a lot more specific right now are the ones that are going to win because everyone's still thinking about doing content for the SEO world. They're still producing very highle top offunnel content. They're not even really focusing on painpoint SEO yet. But what we're saying is now content strategy needs to be even more specific than just targeting bottom of the funnel keywords. You have to think

about all these different scenarios and write for those. Welcome to the Grow and Convert Marketing Show. I'm Davish. That's Benji. Today we are going to talk about how a lot of our AI search and GEO thinking and the different articles we've published and what we've learned from our client work from building tracker and tracking this is starting to coalesce into this strategy or framework that we're calling topicbased GEO which I know is a game we've used before, but I think we're now starting to get our mind wrapped around what the framework and strategy should be. So, we're kind of going to retake that name, but the leadup to it that I want to talk about is our most recent article. So, just for everyone listening, if you haven't followed every single thing we've done, first of all, why not? Second of all, I will recap for you. So our core framework to date has been prioritized GEO that argued that you should prioritize content that teaches LLMs about your product, what it does, why it's good already. That has the beginning seeds of what we're going to spend most of this episode talking about of topic based GEO. But a lot of prioritized GEO was talking about these three tiers. That was tier one, like your own content that influences and teaches LLMs about your brand and product. Tier two was then getting mentions in other people's content. And we argued a bunch in that post about how you should actually not focus on what at that time and honestly still kind of today a lot of people in the marketing world online are talking about, which is tier three, which we say is like on-site tweaks, changing headers to questions, FAQs, LLM, TXT. that is just stuff you do to try to feel productive but it actually doesn't expose your brand to LLM. That still holds very much holds. We've been applying it. Those are kind of the early frameworks of what we're calling topic based GEO. Then the other one that we have is that's key and foundational. It's still very much true is the post you did even before that basically I think the title is something like why ultra specific bottom offunnel content wins in AI search. Don't hold me to that. — I think you're right. — And that one you wrote kind of at the early stages of when this was blowing up in 2025 and last year. — I wrote I wrote it actually in 2023. But essentially the argument Yes. But the argument was that with AI, you're just going to need to be way more specific in terms of what you're writing about just because of how people are searching. So Google was people are primarily searching like short form. They're searching head keywords. They're just searching like short phrases. And so Google tried basically train people to search that way. And then with AI we saw people using a lot more longer form questions identifying themselves saying I'm the CEO of this company. this is the problem that I'm trying to solve. What is the solution for me? And so when we just saw how people were searching, we came to the conclusion that content was going to be need to be much more specific because a lot of people are writing for keywords and they weren't taking into consideration all this identifying information. And so if someone's saying, "I'm this person. I'm trying to solve this very specific problem. What's going to be best for me? " if you didn't have content that spoke to that really specific question, you weren't going to be surfaced as an answer. And so essentially, we just argued in the future content strategy is going to need to change by instead of writing for these really short form questions, you're going to actually want to be really specific, talking about really specific case studies, use cases of your product, talking to specific personas. Um, and yeah, so that was uh the article that I wrote — and that definitely is the early seeds of what we're going to talk about here with topic based GEO. You had a second argument in there that's important that I need to make sure we say so that the readers are on the same page or listeners page, which is you also made an argument at the beginning. By the way, AI search is kind of killing top offunnel content. That's kind of separate from the specificness specificity that was first of all top off ofunnel content like the how-tos intro guide to your industry blah blah any kind of justformational question chatbt and all will just answer it for the user in fact catchb itself often doesn't even link to sources unless it really needs to so it's not going to mention your brand and like link to you and even if it does if someone's just like asking in our case hey help me build a content strategy G that's you know SEO and GEO friendly for example like we in the Google world we would have an article on that they would Google it and then we would show up they would be on our site and be like oh this is really smart who are these people or whatever that's the normal flow that the top of funnel argument is they're just going to have that conversation with Chacht and it has zero reason to mention grow and convert so that was that now with those things we've said kind of in the books right content on your site be specific top of funnel is dead. Let's continue to the

latest thing we just published and then we'll get to topic based geo. That latest thing we called invisible prompts and it was kind of another extension. So you said 2023 that like blows my mind by the way just how old and how fast time travels and how old we are. So um I swear it was like 6 months ago that you published this. Anyway, so then I think the evolution of that we saw especially working on tracker where we got exposed to more than just our clients but like people who use these visibility apps is I think what a lot of people started doing is like okay they understood that keywords what was SEO keywords is now prompts and so like let's track our visibility with like a longer question which is the start of what you're saying they would use these other tools we did you know we did reconnaissance of other tools peak and profound and like you put in your brand name into one of those and it spits out these questions. Heck, tracker also to some extent works that way although it is topic based. We can talk about that maybe in another video. But that was kind of how the world was operating. You know this marketing GEO world was operating. What invisible prompts says is actually what we've learned over maybe 2025 let's say is even that is insufficient that being switching from a short keyword to a long prompt and be like oh Benji look I did what you said in your article I have like a long question so instead of content marketing agency enter which is the Google SEO short keyword look I'm tracking in my tool help me find the best content marketing agencies for SAS look I put SAS in there. It's like a full question like isn't that enough? — Yeah. And a lot of these other tools that have come out recently have tried to position themselves that way by saying we're taking clickstream data or we're taking data from somewhere and seeing — and we know those questions. — Yeah, we know the actual questions, but knowing the actual questions still doesn't help you in terms of being able to track what you show up for. Uh mainly we published a little bit about this, but the study that Ran Fishkin did recently showing that if you run the same search over and over again, you can see different brands show up for them. And so same thing, even if you found the same long form question, if you were to run that same search over and over again in different browsers, you're going to see different answers show up. — And Ran's study was like, I think he just got a bunch of volunteers to do it. I believe in an incognito window. I'm not sure. I would need to double check that. Now, what Invisible Prompts argues is actually and let me pull up for those that are going to be watching this on YouTube. I'm going to pull up this nice graphic um that we ended up making for this because the graphic itself shows a lot. So in this graphic what it argues is it uses an example of find me a content agency with an iceberg and you kind of see above the water and below like the waterline you know in an ocean is kind of um midway through the image and on the above the waterline and the iceberg that's visible it's just someone typing find me a content agency into chatbt below the waterline we have kind of a schematic or an example a hypothetical of all this detail The idea being chat GPT or any LLM has so much personalized context about the user like what type of business it is, what's their goal, what's their preference, what they've used before, what they've already tried, what they failed. What is that? That's symbolizing the massive amount of context and history that these LLMs have about their users. As AI tools have become more popular, basically every single person that's using them, like the kind of lead that you're looking to get is the lead that's on chatbt and going there instead of Google. First of all, who are those people? They're not people that are searching in anonymized logged out G Chhatbt. They are logged in. They use chatbt for everything or whatever their preference is, Claude, whatever, Gemini. And so these things are developing these long contextual histories. All of these tools now or at least chatb and claude at the time that we're recording this have options to be like remember different things about me across all of my different chats. What I'm saying is even if you are tracking the exact prompt that a user happens to type in even that by the way is low probability because a user is going to have a conversation. Most people don't treat this like Google and type find me a content agency enter. They have a long conversation about their business problems and then eventually that leads to content agency. That's one thing. Even if you get lucky and you type in the exact you're tracking in your tracking tool, the exact prompt that the user is typing in, what Invisible Prompts argues is you're still not going to get the same answer as them because you lack their context. And to say what you said, like these other tools that are saying, "We have clickstream data, so we know what prompts are most popular. " What that's saying is that still doesn't help you because the prompt is just that last thing that the user types in. But what is dictating the answer? That's what you want to track. We want to show up in the chatbt answer. What's in the answer is all this context behind the prompt. That's why we're calling it invisible prompts. That the full prompt includes the history and 90% of that is invisible. All you see is that last little thing they type in. So the argument here is we all as marketers and you or whoever is listening, you can't be thinking about GEO this way and tracking it this way and measuring success this way. What way? like SEO, you can't think of it the way you think of SEO keywords, like are we ranking number one for best content agency? Like you can't do that anymore because that was predictable more or less. Tomorrow it's the same results on Google, but we're saying today it's not even the same results for two different people because they have

their actual history and content agency is going to include all this specific personalized stuff. And I had this really kind of fun example of one of our current active clients who typed in uh her prompt was what are the best SEO companies who by the way who said I found you on chatbt. It kept recommending you I needed to talk to you guys. She's now uh one of our good clients and we said can you for our own education send your chat. So she just like copied and pasted it and her exact prompt was, "What are the best SEO companies in your opinion who have driven massive results with endless leads for people in our kind of space? " And so I was like and the number one result. So it had I have the full start of their answer and it mentions grow and convert number one. Great. So then I was like, how would we possibly track this as marketers as an example? So I could type the exact same thing in. So I went to an incognito window and I typed that exact same thing in. And actually in our kind of space doesn't make sense in incognito. You have no chance of it doing the same thing as her. Which by the way just that shows you an example of what we're talking about for leads for people in our kind of space. You're what are you supposed to track? She typed in our kind of space. She's already talking to chatbt. The prompt itself implies like you know what space I'm in. So, I was like, "Okay, in order for this example to be somewhat realistic, I need to like define that. " And I think I typed in something like in the finance space or whatever um result she's in um business loan space or something. And I typed it in and we were nowhere to be found, much less number one. And I have an example of the start of the answer and it's just totally different. It not only recommends a different agency, like everything about it is different. And then the rest of it, we were just not in there. And so that's the protottypical example of an invisible prompt. And so naturally what this leads to is so what do you do? What's funny is this example

is actually like a positive. In real life it recommended us. We got a qualified lead. We closed the lead. We have a great relationship with this company, right? It's like a great client. And then but then in in marketing tracking world, we're like uh how would we even know whether we're doing a good job and our like strategy like how do we know what worked? Why did Chach recommend us to her in her situation? Cuz we're typing the exact same thing in and it's not even close. So what do you do? And that's what's go down a little bit. Yeah. So the key is in that second part where it says real case studies where they 10x lead volume for B2B for SAS and B2B. This starts to get at the solution a little bit which is what we're going to talk about which is topic based GEO. Topic based GEO is our strategy. We're in the final stages of writing a post on this in terms of how content strategy needs to change from let's say content 1. 0 0 world where it was just SEO to content 2. 0 O world which is SEO plus AI search which is essentially you need to cover every area around a specific topic. And so for example, if we are a content marketing agency that largely services SAS businesses, what are all those questions that a SAS business marketer or business owner would ask about doing content marketing? Well, so from a just trying to find an agency, which is super bottom of the funnel, to case studies that show that we can get results for those types of businesses to all the use cases that they're going to have. So like how to do content strategy for a SAS business, how to measure results what is the ROI of content marketing for a SAS business. What changes is in that 1. 0 0 world where we were just focusing on ranking for keywords. We had the painoint SEO process which was just know their pain points and then target keywords attached to those pain points. But now what changes is again we need to go much more specific. So if we have multiple different types of SAS companies, so if we have enterprise SAS that we work with, mid-market SAS, like a soloreneur running a SAS business, and maybe the content strategy would slightly change for each of those different businesses. Now, we need to delineate that in the content that we're writing in case studies and content strategy. we need to have way more pieces explaining how we do content marketing for all these different types of SAS businesses. Whereas in that content 1. 0 world or just for uh SEO, we could just have SAS content marketing strategy and just write one piece and rank for that and that largely did the work of showing up when people were searching around this topic. Whereas now, again, — the reason why is really important because in that SEO world, all of those different types of customers that you just mentioned, this type of SAS, that type of SAS, they're kind of just typing in the same thing into Google, right? — Yeah. They're not going as specific. — Yeah. In chat GBT or an AI search, even if first of all, they probably don't type the same thing in. They'll have a more customized long prompt, long conversation where they're like, listen, like this is my scenario. I have these problems, blah, blah. And there's the twist I just showed with um that example. I almost said their company name or her name, but is the example is um even if they don't type it in literally, Chachi already knows because they've had previous conversations. And so what you just said was you need to have scenarios for the problems. what why you're using us as an example our business you need to have content around like the pain points of enterprise versus the small business example and then like all these things that they may ask for because both the way the user searches in AI they're more specific and the AI itself even if the user doesn't type it in is so smart it's bringing in all of those specific things so the game becomes think through all these pain points, use cases, scenarios, and we're using us as an example, but for product companies like SA software, all of your little features. When we talk to software companies, they'll be like, "This feature helps this really specific situation that are like customers have those benefits, the use cases, those like solution and industry do drop downs you have on your marketing site. Why? Because you don't know the exact prompt they're typing in. you don't have this keyword research tool and as we argued when profound or whoever says we have clickstream data it it's not that meaningful because they don't actually know the personalized background. So what you have to do is you have to say I know we're talking to our customers just like you said Benji about us. We know we have enterprise SAS versus mid-market versus SMB versus soloreneur and we've talked to them and we know their unique pain points. The enterprise SAS company has very different pain points for why they need a content agency and what they need out of it than the soloreneur. And we also know various situations or problems that those people would face inside their company as it relates to content marketing. So for example, they've tried other agencies and those agencies didn't do well for various reasons and going into the specifics of those reasons and then they're looking for another agency. they've tried it in-house and it didn't work for these various reasons and now they're looking for an agency and so — they have a blog and it has a bunch of traffic but it doesn't convert versus the younger startup that's like we actually haven't even started this channel. We don't have a blog — and so when we need when we're writing content talking about all those scenarios in our content is what helps us be surfaced. And so even you saw in that answer how it said this company has a number of case studies showing that they know how to do it. That's the proof. So like one you have to be able to talk about this strategy in detail or how you guys do something differently than everyone else. And I will say that the companies that are being a lot more specific right now are the ones that are going to win because everyone's still thinking about doing content for the SEO world. everyone is still thinking about. They're still producing very highlevel top offunnel content. They're not even really focusing on painoint SEO yet. If you're just doing painoint SEO, you're already doing better than a lot of the other companies out there. But what we're saying is now content strategy needs to be even more specific than just targeting bottom of the funnel keywords. You have to think about all these different scenarios and write for those and produce a lot more content talking about all these different scenarios and situations that you guys do something differently. So when the LLM is going to search and they're trying to match which company would be best at solving this problem for the user that I'm trying to answer this question for. If they have a ton of proof on their site showing how they've solved this problem for this specific user and they look across the web and no one else has anything talking about how they've solved this, they might talk about it at a high level in one of the posts saying, "We're great for some enterprise SAS company, but then there's no proof to actually back it up saying, "We've actually done this for this company, and this company, and here's how we did it specifically. " They're not going to recommend you because the LLM is trying to not They don't have the information to recommend you. They don't know that you're a fit for it. And the example that you were just talking about in our example was here you your eyes noticed two things that in her prompt it was what are the best SEO companies in Europe who've driven massive result with endless leads for people in our kind of space. You're right. Her prompt kind of is begging for a case study. It's saying like give me companies with this proof for quote our kind of space and Chachi PT knows everything about her clearly because it's responds with great question especially for a business funding company like yours where look at this look at how smart that is it knows like what it needs to do where ranking for high intent keywords eg business loans fast working capital MCA for contractors can drive serious deal flow — yes it already has — what I what I see in this is it's picking up on our strength strategy differences by talking about painoint SEO where it kind of hints at these are the types of keywords that you would want to focus on because you're saying you need to drive a lot of leads and so painpoint SEO — talks about this framework in detail where we're saying you need to focus on specific types of keywords if you want to focus on driving results and then if we get into the strength it goes into our specific case studies I already know it's pulling from the geek geekbot case study when it says 10x lead volume. — Yeah. — For these types of businesses. So again, having a broad — which is crazy by the way because Geekbot is like a SAS product and as you can tell here that this client is like in business in the business loan space. — Yeah, it's different but again it's it's trying to it's searching the web and even searching different agency sites and trying to figure out the best match for this company, — right? and it and we have a breath of case studies that talk about how to do content strategy for SAS businesses and how we've done it successfully. And then when it goes to why it works for MCA, this is also important. They understand how to create pages that rank and convert, especially for niche financial verticals. And so it's also talking about our ability to write very like quality content for a very specific type of business and audience. And so it's the combination of everything that we've written on our site that is leading to this answer. What's crazy is we didn't write any of that to rank for this. Like we didn't write it for this GEO purpose. That's kind of the main argument of invisible prompts and that's the leadin to topic based GEO. That's the switch in SEO. It's like you pick the keyword, you write content to rank for that keyword and it does if you do it right. And here it's like your Geekbot case study obviously wasn't written with her in mind. It was written like a million years ago and it has it's not even in her space. It's like SAS and it says that. — Yeah. And but our case study strategy and a lot of the stuff that we've written in case studies was never focused on SEO actually. — No. What we were trying to do is — with every client that we wrote a case study about, we were trying to talk about a very specific situation that we helped solve for this company that we knew a lot of our other clients also fit that persona. So, as an example, with this case study that we had for Circuit, they moved from a subdomain to subfolder. We actually just had a client, actually we had two clients that have signed with us in the last year that also were kind of weighing should I move from a subdomain to subfolder? And so talking about that situation in detail about how we made that switch and how it improved that traffic now is super relatable to the same companies that are weighing those decisions in house. Then we have another case study on an all-in-one SAS business about how to do content strategy where you have multiple different feature sets and you're a large business. And then we have another case study where it's like a self-service SAS for smaller businesses and how to do content strategy for there. So again, when we're thinking about writing these case studies, we're thinking about all the different situations that our customers find themselves in and writing a story around that specific story arc. So that now when people are searching an AI search, when they're identifying themselves, I have this specific problem in our business. Do you know someone else who can solve this problem for me? We have a ton of this proof on our site that shows the LLMs that we know how to solve these problems for all these different types of businesses. And we need to continue to do an even better job at this. Now, now that we're seeing this personalization get taken into account, we need to think through our current client list and think through what are the specific story angles that we need to answer to help continue to build that breath of knowledge um in the SAS space. And same thing even from like what are all the different articles that we haven't really written about SAS yet that we need to — I mean do you know to peak behind the curtain one thing I'll immediately say is we're literally working like an hour before we recorded this I was editing a case study on AI search and like GEO results we had for a particular client because all of our past case studies obviously historically are for regular SEO results and so we're sort of building that up so that some when someone has a similar conversation as she did with Chachib asking like give me examples of agencies that have really good track record in getting their clients on Chachubt or AI search. We have that case study sort of um catalog for the LLM to pull from. So perfect transition now into topic based go which is this thing that's sort of in progress. What we what I said at the beginning was this is kind of the culmination or coalesing of all of these different concepts. I want to walk through this little draft we have, but let me te tee it up this way. You could be listening and be like, "Okay, Benji, fine. " Like, you don't have to I'm not supposed to be targeting specific keywords. We don't even know what prompts there are. They're invisible. Whatever. So, what you're describing feels like what should I do? Just like create a ton of content. Like, is it random? And this topic based GEO concept and strategy is meant to kind of answer that and provide guidance to it that is not just like, "Yeah, yeah, just create a bunch of content like randomly. " — Yeah. Because I do think that's what a lot of people are doing. They're seeing AI be able to produce content a lot faster. So there's like, yeah, let's just like write on every single topic. But no. — Oh yeah. — That that's still going to lead to the same problems that you had before where — people were trying to just write on every top ofunnel topic and then bottom of funnel topic and they were just trying to like produce a ton of content and it was just kind of random and halfhazard. were saying, "No, there's still strategy to this. " And prioritize GEO kind of got as at what tactics should you prioritize in your content strategy. And we argued you need to focus on producing content that goes into detail about your product. And then also the second tier is getting your site mentioned for those topics on other sites. And now topic based GEO is really meant to help you think through just similar to painpoint SEO and what it did for SEO keyword strategy focus on these

three specific topic areas. Topic based GEO is meant to help you think through what type of content do I need to produce that to help me show up in all these different LLM. — So here's kind of the background. I'm sharing this Google doc. A lot of this is work in progress, but we thought it'd be fun like on the YouTube channel here to kind of look at us in draft uh stage. So, the working title I have that's really a mouthful and wordy right now is topic based GEO, creating detailed content about your product and its use cases for LLMs. What I want to focus in on, so I we walked through a lot of the arguments that we just walked through on this episode. So, I'm going to skip over that. I want to move into and you can see here there's even some old phrasing that I'm just gonna literally write on the episode fix so that you guys can see us kind of working in real time. But we initially called it like having a product detail library. I'm starting to home in on this phrase like topic map. And so let's start with the this part. Let's or let's end with this part which is the kind of four or five steps here that we're wrapping our minds around or kind of starting to to zero in on for what topic based GIO means in practice. Number one, and this is kind of the critical part I wrote, we map every product angle that matters. So that's what you should do. Like you need to every brand, everyone listening, if you want to do this, you need to think about every single angle like and and I'll read a couple of these examples. Your categories, your competitors, your use cases, the target personas, the jobs to be done, the pain points it solves, and that's your target topic map. I love this line. The topic map is the full universe of conversations where an LLM could potentially recommend you. That's what Benji you were talking about is like now we're in this scenario where you don't know like SEO beforehand what people are searching with like oh this keyword has this volume but what you do know is you do know your customer or you should if you don't you have other problems geo is not your problem you have other marketing problems or business problems so you need to know your customer when you know your customer you know the questions they're asking like listen to the sales conversations which by the way like we've been saying that for 10 years. So it actually is coming full circle. And as an aside, I've had conversations with other folks in SEO space who say LLMs and AI search, although it's disrupting our whole industry and making us think through all this stuff. It's kind of a lot of people are arguing it's pushing everyone to be better marketers. You can't hack it the way you can hack Google where you'd have some big brand that produced kind of crappy content that just always was on the first page because their domain authority was high. Like now the LLMs are so smart. They're not just like, I'm just going to pick the biggest domain authority for this like head keyword. That doesn't exist anymore in AI search. They're like, oh, you have this really specific problem. Oh, you need business loan like a content agency to 10x your leads. Like let me go through and figure out who that is. And so your topic map is you kind of laying the seeds on the internet with content that teaches LLMs when they crawl, when they search the web, you know, when you type something into chatbt and it says searching the web, you need to be there when your customers doing that and having a similar conversation as we just showed you for us of all these different use cases, scenarios. You need content that the LLM discovers and says, "Oh, that exactly is like what my user is asking me right now. this company, maybe I should recommend them. — Yeah. I think a key thing is that I feel like a lot of people are focusing on like the hacks and tactics and the things to drive short-term results. If you don't think that the LMS are going to continue to get smarter, just think about the evolution of — LMS from 3 years ago to now and how much better they are at both writing and finding answers for you. That's going to continue to improve. And so there's a lot of people that are just focusing on like I don't know putting a bunch of Reddit information on spamming social media like schema like trying to trick the LMS into — answering the question. I think the proof part is going to be way more important than most people think because now it can scan your site and see if you have an exact case study that matches this person's like what they're looking for. And so if you don't have that proof point, like I feel like what a lot of people are trying to say is like even with list posts, they're trying to say I'm the best at this — top 10 content marketing agencies. We're just going to produce a bunch of these and then say we're the best and see if it ranks us. That's not going to work in the future. It might work now, but it's not going to work in the future. And if you're going to say that, you're going to need proof that backs it up. case studies that show you can actually execute on the service offering that you say you're going to. You're going to need other people also saying that this content marketing agency is really wellknown and has smart thoughts and they do what they say they do. They're going to look at review sites. a map of all this different stuff about your business to determine if they're going to show you or not. And so yeah, I think you need to start on this now in terms of thinking through what all those topics are and then also thinking through what type of proof that you need to put on the internet, whether it's getting reviews from your clients, whether it's producing case studies, whatever it is that proves whatever you're saying publicly is true. — Yeah. So then uh step two is how do you kind of figure out what that map is? What all those different scenarios and specifics are? We say step two you interview the people who know your customer best. We often have said

internally like we interview customerf facing folks in your organization when we kick off a new engagement with any client. If there's a sales team then it's sales. If it's some self-s served product type thing, then it'll be like customer support plus the founders plus whoever has interacted. Why? Because you that's where you get the specific stories. It's the sales team that'll tell you like, yeah, this type of customer comes in and they have these really specific questions or whatever and if they ask these questions, we know they're a fit. You need content on that. And you're operating like a little bit blind in GEO compared to SEO. Like you're not like, oh, there's a keyword for this, but you know, we'll get to that. But you that's what you need to write in whether it be in case study form or like how-to form. And then what we do next and this is important. So step one and two is like how do you figure out these topics, right? That's what we've talked about today. Step three on the screen that I'm sharing says we produce content for each angle at the depth of the sales conversation. I'm probably going to tweak that wording when we turn this into a final draft. But what that means is now in the piece you pick your topics. Now for each piece you need to discuss your product at a really indepth level like you are on a sales call demoing it product service whatever. This is another shift from old school content marketing. Old school content marketing was very shy about being quote unquote salesy. it was like considered you have to add value and that's why it was all this like top offunnel how-to stuff and you didn't want to like be too salesy about your product. You kind of like inserted it in certain places. That may work a little bit, but here we're talking about all these use cases and scenarios where your product shines. You need to talk about your product in great depth. Like if we go back or just remember that example we showed, it was like they've 10xed leads. They do bottom offunnel stuff and that applied to her scenario that she was asking chatbt about like this would be a good agency for you. That's what you need to do. So you're So actually when you mentioned list posts I'm kind of like it's like a pet peeve of mine that there's this narrative going around about like list posts aren't good. We're working on like some data to show otherwise. List posts are killing it for our clients and for us. Why? Because our list posts aren't some AI slop nonsense that has this little paragraph about everyone, including us. Our list post, whether it's for us or our clients, has pages of detailed screenshots, examples, features, use cases for the brand that it's for, like our client or us in our case. So you're having this extreme you're providing for the LLMs this great body of what do I want to call it documentation like content of all of the details in which your product works so that when these really specific personalized conversation happens the LLM have that to work from and have a chance of recommending you. And then the two last parts that we're probably going to need like the last part will probably need another video to talk about it. But next after you kind of think through this you can where appropriate pair this content with a Google keyword and there's a lot of that overlap and that's how SEO and GEO has some level of ven diagram overlap is if you are writing about what did you say for us like enterprise like content marketing scenarios and how we work for enterprise there probably is an SEO keyword in fact I think we have a piece on this recently that we published and is already ranking enterprise content marketing And it's totally fine. We find it's still effective in our data to then target that for enterprise content marketing. You'll rank in SEO for it, which like reminder to everyone, people still Google things and you still get SEO leads. Like we're like acting as though that doesn't exist anymore. It does, but then you have this body of content for the LLM to also discover when they search the web that has the spec specificity. And then this we have an old video that we had initially called or an old episode called topic based geo. We're going to rename that probably to like topic based visibility. But the last part is you then need as a marketing team — to track your visibility at a topic basis. — That episode that article we've published that we're going to rename topic based visibility talks about that. So I don't want to get into too much depth here, but like tracker is built in this way, which just to pat our own back, we kind of had that inkling many months ago when we decided to sort of rebuild tracker that like single prompt basis was nonsense. We're like, what is this? It's like just it's like — because we init we initially had built it the wrong way. So tracker — and everyone else did. — Yeah. Tracker taught us a lot about just how AI search worked. We initially approached it by just like on a keyword basis. We were just bringing in SEO keywords thinking that that's how people would want to track here thinking that the keywords were a good proxy for that more personalized search. But then we realized that wasn't the best way to approach it and we came up with this topic basis to run. you have a topic and then you have a bunch of different props prompts below it that get at that same topic. And so the more times you show up under that topic area, the more likely you are to show up no matter what personalization or what the person types in around that topic area. So that's currently how the product works. But yeah, I think to tie this all together, we've had a lot of these disperate ideas about how AI search kind of works over the years. the specificity, the topic basis, why you shouldn't measure individual prompts and — on content, not hacks. — Yeah, exactly. Yeah. The on-site tweaks don't matter as much as people think they do, that argument really never made

sense to us. Because if LLMs crawl the entire web and are smart enough to distill all this information to really smart answers, why would you need some certain schema on your site or some certain format or some — you got to change your site to markdown because that's — all these different formatting things to just read content. They can already read content. That's the whole point of them. It just never made sense. And for whatever reason, everyone in the industry gravitated towards this stuff because in my opinion, they could sell something. They could just they could take all of your content and change the formatting of it and sell you that — people love hacks. It's like just like human nature. Like we're just like lazy. You want some like quote unquote one weird trick. You're like, "Okay, we're going to add an FAQ and that's going to solve our GEO issues. " What we're talking about is hard, by the way. Like we're talking about like real this is what I'm saying is like I've had these private conversations where people are like you know what this is just making everyone have to do good marketing like we're talking about like okay figure out all these different use cases and scenarios that your customers may be like asking chatbt for and then write really detailed content about what about the ones where you think you shine and why you shine and how you shine. That's just like real marketing. It's not a hack, right? And in fact, in invisible prompts, I have this section at the bottom where I say it's kind of I don't know if it stayed in the piece when we published where this is kind of like word of mouth. It's like modern-day word of mouth because these conversations before AI just happened between people. — Exactly. And now and now people are going to the LM as their trusted confidant and asking these questions to the LLM. And so you want the LLM to recommend you and the way how is you need to provide it enough context so that when people are searching for really specific things, — it knows that you even do this thing. And then you need to back it up with proof just like a a person would. Person wouldn't recommend an agency that they've never really worked with or a product that they've never used. And it's the same thing here. You need to have proof online that your product or service does what you say it does and that's how it's going to help recommend you. I think we were hesitant to jump the gun on a lot of AI search strategy until we had felt like we've tested enough things and figured it out enough on our own to where we felt confident like expressing our framework and thinking around how to do this and we feel like the topic based GEO is equivalent to when we discovered painoint SEO. We feel pretty confident that now having done this across a number of clients and seeing from clients who have never done marketing before starting to see a ton of visibility across the LLMs that this approach is working. And so now that we're going to launch this new framework of topic based GEO, we're going to continue to share case studies and proof that this works over the coming months uh to just back up this thesis. But we think we're really on to something here with covering your basis around a topic area and having that help you improve visibility across all the different LLM. If you like this video, don't forget to subscribe. You can also get the audio only versions of these shows wherever you get your podcasts. And you can follow us at growandconvert. com/newsletter for any articles and updates for when these videos come