What does AI retrieval behavior mean in this evidence set?
Source-backed creator statements and evidence excerpts related to AI retrieval behavior.
Topic evidence page
Source-backed creator statements and evidence excerpts related to AI retrieval behavior.
Source-backed creator statements and evidence excerpts related to AI retrieval behavior.
AI search often breaks a user's prompt into several specific, less competitive searches rather than one broad keyword query.; The source says AI search may run many specific searches, inspect multiple result pages, and consider hundreds of pages before recommending.
AI search often breaks a user's prompt into several specific, less competitive searches rather than one broad keyword query.
This topic currently has 2 source records, 2 public insight cards, and 1 creators in the public Base2026 export.
it very clear. And so if we're moving into a world where all search is AI search, then that's what SEO is. SEO is all the things you do to get recommended in AI search. We don't need a separate term for that. If you're doing SEO and you aren't focused on AI search, I would say you aren't doing SEO, at least not for much longer...
Openoing a single broad search like dentist near me the way a human might, it'll do up to a dozen very specific searches. And then instead of just looking through the top few results like a human might, it'll look through multiple pages of search results...
OpenWhen people hear that I do SEO, they often ask, what about GEO? Do you also do GEO? Or people will come to me and they'll say, I already have someone to do my SEO. I'm looking for someone to do GEO or AI optimization. And I always say, sure, I do what you're referring to as GEO, but I just call it SEO...
OpenTraditional search engines are going to be dead within a few years. I wanna talk about how we know this and then who the winners and losers are gonna be in this shift. And then finally, I'll talk about how you can make sure you're one of the winners...
OpenShort public snippets grouped with their source record, creator, and date.
When people hear that I do SEO, they often ask, what about GEO? Do you also do GEO?
Or people will come to me and they'll say, I already have someone to do my SEO. I'm looking for someone to do GEO or AI optimization.
And I always say, sure, I do what you're referring to as GEO, but I just call it SEO. And my guess is within a year, we'll all just be calling it SEO.
But that's not because nothing has changed. You'll hear a lot of SEO saying that the strategies for ranking in AI search are exactly the same as the strategies for ranking in traditional search, that by doing traditional SEO, you're doing everything necessary for GEO.
These are people who are not keeping up with SEO. So I'm gonna explain exactly how it's changed.
But first, why don't I like the term GEO? The term GEO implies that traditional SEO is still a thing, and now there's this new thing you also need to do called GEO.
However, from my perspective, the way that we search for and find information is changing, and SEO just needs to change along with it...
it very clear. And so if we're moving into a world where all search is AI search, then that's what SEO is.
SEO is all the things you do to get recommended in AI search. We don't need a separate term for that.
If you're doing SEO and you aren't focused on AI search, I would say you aren't doing SEO, at least not for much longer. You're doing traditional SEO or legacy SEO.
It's similar, sure, but it would be like someone in 2020 using SEO tactics from 2005 and calling it SEO. Okay, but what do you mean, TJ?
How is it different? I have a lot of videos going through the differences in details, but I'll give you the gist.
In traditional SEO, a user is interacting directly with the search engine. They're typing in a query, and an algorithm is returning about 10 results.
The user is then going to click on one or two of those websites before they get the information they need to make a decision. Typically, they're only going to look at the top three results.
It was a big decision. They might do a few follow up searches, but typically it's just the one...
websites in the top 20 results of each of those searches. For that reason, ranking in AI search is actually really easy.
As long as you have an ounce of authority and you have content that's somewhat relevant for one of those dozen searches, you can easily be one of the hundreds of pages that surfaced. But ranking's not what matters in AI search.
It's getting recommended. Of the hundreds of pages that the AI reviews, it's only going to site a few of them.
It's gonna try to determine which of these pages are most trustworthy and relevant to the user's query. And it's gonna retrieve not the entire page, but just one passage from that page and pull it back into the context window of the conversation.
And ultimately, the only thing that matters is that your brand is recommended as a solution in that passage. That's what SEO is now, or at least what it will be very soon.
If you aren't doing that, you aren't doing SEO, at least not for much longer.
Traditional search engines are going to be dead within a few years. I wanna talk about how we know this and then who the winners and losers are gonna be in this shift.
And then finally, I'll talk about how you can make sure you're one of the winners. Kevin Indig recently reported that Sundar Pichai, the CEO of Google, says that he sees Google's AI mode replacing traditional search within two to three years.
So that means what we now think of as Google Search would just be a separate tab that people can go to, but the default experience would be AI mode, which is a lot like ChatGPT. However, the shift from traditional search to AI mode has already begun.
Google is already experimenting with bringing more and more AI features into the traditional search results page. So it seems like it could happen much faster.
So what is this going to mean for websites and small businesses when suddenly people are going to AI to get their answers instead of traditional search engines? If you're someone that depends on web traffic like a publisher, then this is just bad news...
ryone kind of accepted that as common wisdom back then. You couldn't trust anything on the internet.
You had to drive to the library, find a book, bring it home, and read through it to find information. Nowadays, if someone says they research something, it's a pretty safe bet that they went to Google, they did a handful of searches, they clicked on some of the links, and they read information from the websites.
For about 10 or 20 years, this has been standard practice before making a big decision. That's all about to change.
iPoll Rank did a study that found that only 4.5% of searches in Google's AI mode result in a single click. I think within a few years when we hear someone say I researched something, we'll be able to safely assume that they went to ChatGPT or Google AI mode and typed in one or two prompts.
And I'm not passing judgment. I think as these large language models get better, that will be the best way to do research.
But what that means is we're effectively outsourcing our decision making to the large language models...
oing a single broad search like dentist near me the way a human might, it'll do up to a dozen very specific searches. And then instead of just looking through the top few results like a human might, it'll look through multiple pages of search results.
What this means is that the large language models are considering hundreds of pages before returning a recommendation. And so it's no longer about ranking in position 1 for the most competitive search term.
It's about showing up in as many places as possible. So the brands that are gonna win from all this are the noisy ones, the ones that are putting out content everywhere they can, that are joining the discussions wherever they're happening.
So that's exactly what we're doing for our clients. We're tracking where all these large language models are looking, and we're making sure our clients are represented on as many of the web pages as possible.
By the way, if you're interested in keeping up with this stuff, Kevin Indig's newsletter, Growth Memo, is one of the best out there.