815 | What AI Search Actually Cares About (GEO & AI SEO 2026) | Chris Long

You can rank number one in Google but still be completely invisible to chat GPT. That’s Chris Long, one of the most influential voices in AI-driven SEO and co-founder of Nectiv, the team building some of the most aggressive AI search tracking datasets in the SEO world. He’s extracted 60,000-plus fan-out queries from Gemini to see how AI mode actually thinks. Analyzed 8,500 chat GPT searches across every major industry to map what the model really searches for. He’s torn apart Google’s Enterprise AI stack to understand the signals behind ranking and selection and he’s found something that should make every marketer nervous.

We’ve seen AI agents completely ignore large enterprise companies and choose no-name companies with very little organic traffic. We traced AI answers back to their sources and this is what we found. In this episode, you’ll learn how AI engines actually read and search the web and decide what to cite. GPT isn’t crawling your whole page. Every window matters. Every structured block matters. This is one of the strongest signals LLMs are using. How to redesign your content for AI native rankings and guarantee a spot in the AI generated answer. We found that tables were two times more common over two times more common in AI engines than Google search. And how to influence AI search when you don’t even control the pages that matter. You have to know citations that are influencing chat GPT.

One thing that I think is clear is that the AI engines are reading pages differently or they’re rewarding pages differently than Google necessarily. So maybe you could talk a little bit about how you see the AI engines reading a web page. Yeah, so AI engines I think is a fundamental kind of misconception in like the marketing community where it’s like a lot of people think that an AI engine is going to read your page the same way as Google, right? Like Google is going to go, hey, I’m on your best accounting software page. I’m going to read the introduction that I’m going to read the main sections. I’m going to read the FAQs and the referrals and testimonials and all that stuff you have on your page. AI engines don’t work like that. Now there’s not a dedicated, hey, web guide where it’s like this is the way all AI engines are reading the pages, but there’s a lot of research being done, right? And Dan Petrovic is one of the great AI researchers, especially when it comes to digital marketing, highly recommend checking him out. He wrote a great article on his site where he talked about basically how AI engines read content

differently and they basically use different context windows, right? So an AI engine might read only a portion of your page, right? Maybe it reads a fourth of your page right at the top, and then it reads another section down at the bottom. It basically can come in at random code lines and look at your page in these different kind of like windows or chunks. And that’s what oftentimes we’re kind of referred to when they’re referring to as AI is chunking the content. Basically, AI is reading in different contexts windows your page. What does that mean for SEOs, right? And what does that mean for visibility? What means that you have to account for any potential section that AI might be hitting, you having a well summarized answer for what that query is, right? So like you want to be sure even if the AI comes in on the first part of your page or the middle part or the third part, right, you have different content formats that can give it the answer it needs. So a couple of different features that make kind of like the ideal optimize LLM for LLM’s page, right? One, we actually did a study for our company

Nectiv and we actually I compared AI engines to like AI engines to Google search and like the content that the content on pages that the AI site to AI engines were referencing to what Google search was referencing. And we found that tables were two times more common over two times more common in AI engines than Google search. So that content in that really structured format really matters, right? We’re very, very commonly seen that and more people even adopting like the inclusion of tables in their content, because it’s a really nice way to structure your content for engines like chat GPT bullet points are good. And then heading and then your traditional like heading paragraph structure, but like very defined, like concise heading and structure like the best small business accounting software of 2026 is also tends to be formatted well for AI engines. If anyone is like, Hey, I want a case study on this, like who’s doing this great zappier, especially in the tech industry is like the number one example. It’s like who I’m always pointing clients or potential like prospects to they didn’t do it intentionally. I think they

already had it built up beforehand. But like if you go on any of their pages, right, like the best CRM to 2026, like they’ll have at the top a bulleted list in the middle, heading, heading tags, and then like individual like tech sections with bulleted list that list pros and cons of each technology and at the bottom a table, right? And that’s really one of the ideal ways to structure your content because you’re giving it multiple pathways to connect it back to what we were talking about. Now if an AI comes in and looks at an individual context window, maybe just captures the bulleted list at the top of the page, maybe captures the table at the bottom where you’re giving yourself multiple shots on net of structured content, no matter where the AI engine comes in and starts to actually analyze the pages content.

And they’re not only just reading the content differently, right? They also fundamentally search differently than people in other engines. So I don’t know if you have some insights into like the fan out research and what that’s showing, but maybe you could talk about that a little bit. Right now, when you perform a Google search, let’s go back to our previous example, best accounting software, it’s just one search, right? You search best accounting software, Google says, okay, I’m going to check my index and see, okay, what are all the content I have in there, these pages seem good, I’ll filter out these pages, play out yada yada, and then all of a sudden it comes up with the results, which would be your traditional search result, right? AI engines are more complex than that, it actually varies by engine. We’ve done a lot of research of an RAI tracker, my co-founder actually built out the capabilities to extract what are called fan out queries.

AI engines don’t search just that one query, they might search best accounting software, an AI engine might that actually transform that into up to nine different queries, right? So an AI engine might take that and say, I’m going to search best accounting software, but I’m also going to search how to choose the best accounting software, QuickBooks first, FreshBooks. What are the best accounting software of 2026 reviews? It might search those nine different searches, then aggregate the results from all of them, and then return it in the response, right? So it’s actually super charging the number of the complexity of your search landscape, because for every one prompt or query that you want to track, you might have to look at up to nine different queries. In our studies, we actually found for Google, they average about nine different fan out queries, they can go up to 22 different fan out queries. Chat GPT is much smaller. Generally, they search around different, they sort of use about two different fan out queries. So it’s much less complex. The other note on this is that not only are they searching multiple queries,

the queries they’re searching are much longer than something than someone would ever search. So a study by Sierra Interactive actually found when they looked at the Google fan out queries, 95% of them had zero search volume, right? So these queries, these AI engines are searching are not like how you and I would search, right? So that’s something you also have to take into account. And knowing your fan out queries is going to be really important for AI search, because you need to know not just how the user searches, but then how does the AI actually translate that and then transform that into multiple searches. And then you have the content that actually maps back to up to nine different searches per every prompt that you want to be visible for really what businesses should be doing because it’s almost, it’s one almost impossible to account for all of the different combinations of queries.

Like in traditional SEO, you could be like, okay, here’s our list of 500 different keywords. And we have it across our different product lines and services. We’re going to track that and be fairly confident that that’s the, you know, experienced people are getting with these, right? With like, you know, a list of 500, you know, could be a list of, you know, five, you know, 5,000, 10,000, whatever, it can quickly transforms once you do that at scale, right? Actually, it could be up to 50,000, right? So anyways, what I recommend businesses do is look at common patterns across the different AI engines, right? So it’s like, you want to get a good sense across a large dataset, dataset, what are the common themes that AI engines are searching? So a good example is they search for reviews, like all the time. So that means you want to ensure that you have your own reviews page, you might even compare it, you might even compare your product, let’s say your QuickBooks, you compare it against FreshBooks, you compare it across other accounting softwares and have reviews, you need to, and you also need to ensure that one thing we’re seeing is like AI engines search for

like fresh content. So like they’ll search for the year a lot of times, you need to ensure your content’s optimized for that. So instead of trying to figure out all the different, you know, thousands of variations, you need to really distill it and research into what are the common themes, and then how do we optimize our content against those common themes in the future? I actually performed a test at my previous company where it was like, hey, we took a page, we restructured it for more LLMs or we added like bulleted lists, we have bulleted lists, structured out the hide-ins more optimized and better.

And then what we saw is like the page went up in terms of its like SEO presence and then also got cited by the LLMs. So like that’s a good example of like content structuring and actually impacting both. But a lot of those things is going to be aggressive use of bullets, heading subheadings, and then actually integrating tables based on that research. Like those are great way to structure your content. Maybe you could talk a little bit about how to identify like third-party sources or third-party pages that you want to go after to get your brand mentioned in. Yeah. So there’s a couple in this process is going to involve to start, you this would assume you’re using some sort of AI tracker, right? And that’s my number one recommendation for most businesses. If you’re not on one, find it, and you want to be doing more with getting visible in LLMs, you need to have some type of AI tracker. They are not perfect. There is a lot of flaws to them, right? But having some data, even directional data is better than absolutely nothing, right? So that is one of the first things you have to get is an AI tracker like profound, like an Athena peek scrunch, whatever we have our own.

So then what you have to do is once you have, let’s say a list of, to keep it simple, 100 different prompts you’re tracking, you might say, okay, I’m not going to mention for these prompts, right? Well, what are the citations showing up? Like what actual source is this chat GPT listing? And you might say, hey, like New York magazine wrote this article about the five best CRMs or Zapier wrote this article about the five best CRMs, like, we’re not in it. So what you can do is you can actually take the list of citations, right? And then you can actually run it through a tool like Screaming Frog. There’s some different regex you can use actually post about that on my LinkedIn and you can actually insert your brand in that regex and then see, hey, after you perform a crawl, is my brand actually listed in these results, right? In these citations. And if it’s not, then you can start to say, Hey, affiliate team, Hey PR team, should we be reaching out to them?

Is there anything we can do on the own side? There are strategies you can actually use to track your citations and then actually run those citations through something like Screaming Frog to see is your brand getting included? A lot of times some of these tools, I believe Profound does this will actually also automatically do that analysis for you as well. But the key point here is you have to know the citations that are influencing chat GPT, especially for core prompts, and then is your brand actually present in those? And if it’s not, that’s a high priority, that’s a high priority citation to go after. If it’s looking for reviews, like in your local business, for example, that might be your Yelp page, or if you’re a software, it actually might be your G2 page, or a capterra page. So then that’s kind of an almost a reputation management issue where it’s like, Hey, should we be doing things to influence the reviews we’re getting on these third party marketplaces? Because we know that chat GPT, Google are transforming and looking for reviews about our brand, ours aren’t very good, we need to go and get more. So that’s one strategy, right? The other might be an affiliate

strategy where it’s like, Hey, we want to be one of the best CRMs, but you know, TechRadar with this article about us, and we are not included on this, do we have an affiliate relationship with them? If so, then you might be able to kind of talk about an potential list and deal in terms of how you can get structured higher up in that individual article. So all depends on the actual source that’s citing you. Sometimes it’s actually more of an own source than it is, maybe it’s your own listing on Yelp, if you actually control, but it’s a third party site. Other times, if it’s an affiliate, you might have to work with that individual provider in order to figure out, Hey, how do I actually improve my listing and how we’re referenced in this individual article? Talk to me a little bit about what you see as a minimum viable workflow for a business in 2026 to do this work that people could follow in order to understand, Hey, what keywords am I ranking for? Topical things am I working for or am I not being listed? Maybe you could chat about what are the most important experiments to focus on? What people can do is they can use the Ahrefs MCP or SEM Rush MCP. What that will

allow them to do is actually go and access like Ahrefs data, SEM Rush data from their LLM of choice. So what they’ll be able to do is they’ll actually be able to say, Hey, I want to find trendy keywords, chat, GBT around the concept of CRM and then chat, GBT, can then go and query the Ahrefs database and say, Okay, what are trending queries around CRM? Like what’s grown like 100% in the past three months. So you can use it to basically analyze or get access to your SEO data directly in your LLM of choice. So you could start to ask the LLM questions about SEO data. You could go in and say, one common way we’ve done it is gone and say, Okay, here are my top, here are my top competitors, right? With the Ahrefs MCP, what are the biggest content gaps between my content and their content using the top pages report? And then the LLM will be able to access that data and then perform that type of analysis like directly in the LLM without you ever having to log into like the third party provider.

We’re seeing more and more zero click behavior. So people are getting the answer that they need. They don’t need to click over into your website. And so with that, what does winning look like now? And do you have a set of KPIs that you should be sort of tracking and looking at now? Yeah, I think your biggest KPIs shift from traffic may, it might have always been a vanity metric, right? Or a vanity, more of a vanity metric than like a lot of people in the SEO community want to admit, right? Because like you mentioned with the advent of AI overviews, they’re proceeding click through rate go down, chat, GBT, perplexity, like if you look at some of the Cloudflare studies, like their crawl to referral ratio is really low, right? That would look, I think perplexity will crawl something like 30,000 times for every like one session it sends or something crazy like that. Like so in this world, we have to shift to, Hey, what about brand mentions, right? Like is your brand getting mentioned in these AI systems, right? That’s when you have to worry more about things like external citations, right? Where it’s like, Hey, if I ask it what the best marketing or

a best accounting software for small businesses, you know, chat, GBT, perplexity, they might get consensus across a variety of different lists, right? They may look at five different lists, see what the common technology providers are. And then based on that return the results, right? Once again, if you ask it for best accounting software for small business, it might list like, Hey, quick books, fresh books, right? Zero, right? Some of these technologies, that’s, that’s the Google giving it space to be de facto answer. So while you might not get a click from that, maybe someone, you know, does a brand search around you, right? Or maybe someone goes directly to your site. So I think a couple metrics marketers need to start paying more attention to actually think there’s a case to be made for impressions becoming a much bigger factor where it’s like, are you just getting listed in Google search more often, especially for like bottom of the funnel queries, or like highly tied back to your brand? Because like, if you’re getting impressions that there’s a chance that AI overview is actually recommending you, right?

A really popular one that we use in our own AI tracking tool. And if you’re using other AI trackers like profound Athena, scrunch, whatever, like brand mentions in those, right? Like a lot of times that all those tools will track across a given prompts, it how frequently is your brand getting mentioned and also has some sentiment analysis, you’ll also want to be tracking that not can be almost a core KPI across LLMs, right? And then you can say, Hey, am I getting mentioned in chat, GBT? How does that change against perplexity or AI overviews brand mentions? I think are one of the most valuable currencies because even though people aren’t clicking through, they’re still searching for products, they’re still searching for services. Then are the LLMs recommending your brand is like really the million dollar question that I think a lot of marketers are going to need to start answering. Chris, if you look forward into 2026, what is going to separate the businesses that have high success and gain an increasing level of visibility versus the ones that struggle?

Probably the ones who lean into it the most, right? And I think right now we’re kind of at a pivotal moment where a lot of people are saying, Hey, we’re hanging on to the old tactics, right? We’re doing things the same way the same way we used to do things five years ago. I’m not saying we need to completely stop those things. But I think the businesses that thrive are going to very quickly realize like, Hey, things are changing. People are getting more information from chat, GBT. Yes, the studies say that Google’s market share is strong, but there’s also a risk there because they could change their entire experience to an LLM overnight. So the teams that say, Hey, what is our plan for AI? What happens if Google released AI mode tomorrow? What are we doing to account for that? I think are the teams that are going to have the biggest success. So the teams that start early are starting to figure are starting to, when they’re taking a look at like their content strategies, those search strategies, they’re not just looking at Google search, but also LLMs, those are the businesses that are going to be have the most likely success. The other other piece of that I would say is that

LLM search, at least in my experience, tends to favor brands that have already been invested in SEO. So if you have been on the fence about SEO or just haven’t put it like a lot into your content library, like the brands that like are already seen success, like had large content library ones can zap your as a great example. They have this swarth of content or in comparison content that as soon as the LLMs came, like started getting picked up by like chat, GBT and perplexity. So it’s like, though brands that invest in SEO, but also are looking to have the future of search, I think are going to be the ones that see success the fastest.

Pay attention to the changes that Google’s making. And I would say, read the read the tea leaves as much as you can. If you’re trying to, because AI is moving so fast, it feels very overwhelming to myself, certainly. I think one of the best things you can be doing is paying attention, right? Like reading search engine, land search engine journal, understanding like what updates are Google making? Okay, they wrote out an AI mode visibility button onto the homepage. What is that signal about like where they want to go in the future? So pay attention. And then also with everything, every update that happens, try to be thinking critically and saying like, okay, what does that mean for the future of search?

Where, why would they be doing that? Where do they think these things are going that can help you better kind of skate to where the fuck is going, as opposed to just guessing about the future.