What Is AI SEO? A Plain Guide to Ranking in AI Search

Type a question into ChatGPT or Google's AI Overview and you get an answer, not a list of ten blue links. AI SEO is how you make sure that answer mentions you. Here is what it actually means, how AI search works under the hood, and the moves that get you cited, with the free tools to do each one.

Here's the easiest way I can explain it. Think of AI as that one friend who always knows the best place to eat. SEO makes sure your restaurant looks great when your friend looks you up on Google. AI SEO is being so consistently good, in so many places, that when someone asks the friend "where should I eat?", the friend says your name.

Watch: I explain AI search in a few minutes, in the plainest language I can.

What AI SEO Actually Is

AI SEO is the practice of structuring your content so AI search engines can recognise it, trust it, and cite it in the answers they generate. Same goal, three names: you will see it called AI SEO, GEO (Generative Engine Optimisation) and AEO (Answer Engine Optimisation). They all describe the same thing, being the source the AI quotes. We will call it AI SEO.

You'll also see a fourth name, LLMO (large language model optimisation), named after the AI models themselves. If you've ever typed "what is SEO for AI called?", the answer is any of the four. Google's own Search team put it on a slide: "good SEO is good GEO".

The shift is simple. Classic SEO got you ranking in a list of links. AI SEO gets you named inside the answer, often before anyone scrolls to a link at all.

Why SEO Changed So Fast

People stopped typing keywords and started asking questions, and the engines started answering them directly. Google's AI Overviews and AI Mode, ChatGPT search, Perplexity and Gemini now answer the question on the page, pulling from sources they trust. The click is no longer guaranteed, and a zero-click search is now the majority outcome. If you are not in the answer, you may never get the visit.

Two things drive the shift. First, the volume: AI answers now sit on top of a huge share of searches, so the old game of ranking a blue link wins less ground than it did even a year ago. Second, the behaviour: people are asking longer, more conversational questions, the kind that used to take three or four separate searches. The engine handles all of it at once and hands back a single answer. Our guide on how AI Overviews are changing SEO covers what that does to click-through rates (how often people click a result), and why a cited mention can now beat a number-one ranking.

ChatGPT is the biggest of these. It sent 51.58% of AI chatbot visits to websites worldwide in October 2026, with Gemini on 28.86%, according to Statcounter.

How AI Search Actually Works

AI search reads your question, breaks it into a fan of sub-questions, retrieves passages from different pages, and merges them into one answer with citations. That breaking-apart step is called query fan-out: one question becomes many behind the scenes. The page that answers the most of those sub-questions is the one that gets cited.

That fan-out is only half of it. An AI assistant answers from two places: memory, what the model learned from its training data before release, and lookup, the live searches it runs right now. Memory changes over months, so it rewards a brand the web has described clearly for a long time. Lookup changes within days, so it rewards pages that answer the sub-questions today. See the sub-questions for any search with our free Query Fan-Out Tool.

Every answer is also written fresh, so two people asking the same question can get different brands. One test in ChatGPT tells you very little. The same set of questions, checked every month, tells you a lot.

The Truth: AI SEO Is Still SEO

Here is the part nobody selling you an "AEO package" wants to admit. Google has said it plainly: optimising for AI search is optimising for search, and that is still SEO.

Straight from Google's documentation: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." No llms.txt, no AI-specific markup, no separate discipline. As Google frames it, "optimizing for generative AI search is optimizing for the search experience, and thus still SEO."

Google sets one entry requirement: a page must be indexed, meaning stored in Google's database of pages, and eligible to show with a snippet before any AI feature can use it. If your SEO is broken, your AI SEO is broken too.

So ignore anyone selling a magic AI file or a separate playbook. The fundamentals still win. What changed is the scope: you are answering whole questions for an engine that reads passages, not ranking one page for one keyword. You can read Google's own position in its AI features documentation.

How to Do AI SEO (7 Moves)

AI SEO is good SEO pointed at how AI reads and cites. Seven moves do the work.

Lawrence Hitches, who runs AI search at StudioHawk, groups them into three jobs: foundations (moves 1 to 5: can AI read and quote you?), reach (move 6: does the web talk about you?) and proof (move 7: do you know what AI says?). He calls it a tricycle. Take one wheel off and it tips over.

Before you start, take ten minutes. Open ChatGPT, ask the exact question your best customer asks, and screenshot the answer. Date it. That's your starting point, and the thing you check against every month.

1. Be crawlable and readable without JavaScript. Most AI crawlers (the programs AI companies send to read websites) read your raw HTML and never run JS. If your content only appears after the page renders, AI sees a blank page. Check it free with Can AI See Your Page? Then check your robots.txt file lets them in at all with the AI Crawler Access Checker.
2. Cover the whole question, not one keyword. AI fans a query into sub-questions, and the page that answers the most gets cited. Build topical depth and clusters. Map them with the Topical Authority Map and our topical authority guide.
3. Lead with the answer. Answer-first headings and opening sentences, self-contained passages, a real FAQ. AI lifts clean chunks it can quote without editing. More on this in optimising for ChatGPT and Perplexity.
4. Add structured data. Schema is code that labels what's on your page (a product, a price, a business address), so Google and AI can read your entities and pages without guessing. Build it free with the Schema Markup Generator.
5. Add real information gain. AI is brilliant at the consensus and useless at the gap. Add the data, experience or contrarian view only you have, or you are just another page it could write itself. See the information gain guide.
6. Build off-site authority. AI weights brands it sees described consistently across the web: reviews, mentions, citations, profiles. Your entity has to exist beyond your own domain. We did this for Genea Fertility: a data story built from its own anonymised donor records earned 44 pieces of press coverage, and its appearances in Google's AI Overviews went from 89 to 191. Read the case study.
7. Track your AI citations. Rankings alone will not tell you whether AI is quoting you. Track where AI crawls and cites you. The Bing Citation Share report and your server logs show what rankings never will. Run your customers' questions through the AI Visibility Tester too. AI leads move fast: between July and December 2025, StudioHawk's own leads from AI search closed in about 18 days, against 29 for leads from SEO.

Notice what is missing from that list: no special AI files, no markup tricks, no separate playbook. It is good SEO, aimed at how AI reads.

What our data says about where AI traffic lands

We looked at 1.2 million AI referral visits (visits from an AI assistant) across more than 500 businesses in our own analytics data, from January to July 2026. Three findings change what you build first:

• Tools and templates got 7.5 times the AI traffic you'd expect from how many pages they make up. No other page type came close.
• Comparison pages, the page type pushed hardest in AI SEO advice, ranked 5th of 7.
• A third of all AI visits landed on product and collection pages.

Even the address matters. Across more than 500 websites in our analytics data, a page with a readable URL, like /services/roof-repairs/, drew about 13 times more AI visits than a page with a string of random characters. The full numbers are in StudioHawk's study of AI traffic by page type.

The AI SEO Mistakes and Hype to Avoid

Most of what gets sold as AI SEO is noise. Skip these:

• Buying an "AEO" or "GEO" package that is just SEO with a new label. If a vendor cannot tell you what is genuinely different from good SEO, there is nothing different.
• Adding an llms.txt file and expecting magic. Google does not use it to decide what to cite, and has said no special files are needed. Our data from 600+ websites we've studied agrees: is llms.txt worth it?
• Stuffing schema you do not qualify for. Fake FAQ or review markup gets ignored or flagged. Mark up what is genuinely on the page.
• Chasing volume over substance. Thin, AI-written consensus pages are exactly what AI can already generate, so they give it no reason to cite you. Originality is the whole game.
• Optimising for one engine. The fundamentals carry across ChatGPT, Google, Perplexity and the rest. You build for all of them by building well.

The short version: do the seven moves, ignore the snake oil.

Free AI SEO Tools to Do It With

You can run the core checks for free, right now:

• Can AI See Your Page? Does AI see your content in the raw HTML?
• Schema Markup Generator. Build the structured data.
• Internal Link Checker. Find orphan pages and weak internal links so your cluster actually connects.
• AI Crawler Access Checker. Does your robots.txt let GPTBot, ClaudeBot and the rest in?
• Query Fan-Out Tool. See the sub-questions AI search runs for any query.
• AI Visibility Tester. Check whether ChatGPT, Gemini and Claude recommend your business.

And the deeper guides in this cluster: query fan-out, information gain, how AI Overviews are changing SEO, and optimising for ChatGPT and Perplexity.