Query Fan-Out: How AI Turns One Question Into Ten

Ask ChatGPT or Google's AI Mode a real question and it does not run one search. It runs ten. That single move, called query fan-out, is quietly rewriting what it takes to get found. Here is what it is, a worked example you can picture, and exactly how to make your content the source the AI keeps citing. Query fan-out is one piece of the wider AI SEO picture.

What Query Fan-Out Actually Is

Query fan-out is the technique AI search engines use to turn one question into many. When you ask something, the system expands your query into a fan of related, more specific sub-questions, runs them all at once, pulls passages from different pages for each, and merges the lot into a single answer.

The plain version: the old model was one query, one list of links. The new model is one query, a dozen hidden searches, one synthesised answer. You never see the sub-queries. You only see the result.

This is not theory. Google's own AI features documentation describes AI Overviews and AI Mode as "issuing multiple related searches across subtopics and data sources" to build a response, and Google names the technique query fan-out. ChatGPT, Gemini, Perplexity and Microsoft Copilot all run their own version of it. For the wider context on the results page this sits behind, see our guide on how AI Overviews are changing SEO.

A Real Example: Best Running Shoes for Flat Feet

Take one shopping question and watch it split. A buyer asks for the best running shoes for flat feet. Before the AI answers, it fans that single question into seven facets, each becoming its own sub-query.

Now picture your page. If it only covers "best running shoes for flat feet" and ignores overpronation, cushioning, long-distance use, durability and the brand-versus-brand comparison, you turn up for one facet and get left out of the other six. The shop that wrote about all seven is the one the AI cites, because it had an answer ready for every branch of the fan.

Why This Changes Your Content Strategy

The brand that gets cited is the one whose content answers the most sub-questions in the fan. Old SEO optimised a page for a keyword. Fan-out rewards a topic covered as a set of questions. Cover one angle and you are a minor source the AI might mention. Cover the whole fan and you become the page it returns to across every sub-query, the one it pulls into the answer.

That is why AI search has become the new top of the funnel. Prompts, mentions and citations now sit above rankings, impressions and clicks. If you are not in the answer, the old funnel never even starts.

What Your Ranking Is Actually Worth Now

The same position is worth wildly different amounts depending on what sits above it. A classic number one collects most of the clicks. A number one buried under a full AI Overview that does not cite you gets a fraction. A number three that is cited inside the AI Overview is seen before anyone scrolls.

Rankings still matter as the entry ticket. Ahrefs found that in mid-2025, 76% of pages cited in AI Overviews also ranked in Google's top 10. By early 2026 that overlap had fallen to 38%. The direction is the point: being in the top 10 still helps you get cited, but it guarantees less every month. Position alone no longer tells you what your visibility is worth. Being in the answer does.

How to Optimise for Query Fan-Out

First, the good news, straight from Google. You do not need a separate AI playbook. The documentation is blunt: "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 special rewrite. As Google frames it, "optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Query fan-out does not need a new discipline. It needs good SEO pointed at the whole question instead of one keyword.

So stop optimising single pages for single keywords and start covering topics as a set of questions. Five moves do the work.

• Map the fan. List every sub-question a buyer or an AI would ask around your topic. Pull from People Also Ask, and ask ChatGPT or Perplexity to list the questions it would research before answering.
• Build topical depth and clusters. One page cannot answer a whole fan well. A hub page plus supporting pages can. Our topical authority map guide shows how to structure it, and the Topical Authority Map tool builds the map for you.
• Lead with the answer. Answer-first headings and opening sentences. AI lifts clean, self-contained passages, so give it ones it can quote without editing.
• Add a real FAQ. Map genuine questions to the facets in the fan and answer them plainly. This is some of the most quotable content on the page.
• Link the cluster together. Internal links so both the AI and Google can see your pages are one coherent body of work on the topic, not scattered posts.

This is the Relevance pillar of the SOURC-E framework: be the most relevant, most complete answer to the whole question, not one slice of it. To find the facets you are missing, run your topic through the free Content Gap Scanner, then plan the build with the SEO Roadmap.

How to Find Your Own Fan

You can see your fan in two minutes. Open ChatGPT or Perplexity and ask: "If someone asks [your key query], what sub-questions would you research before answering?" The reply is your fan, in plain sight.

Then audit your content against it. Which sub-questions do you answer well, and which do you skip? Every gap is a sub-query a competitor gets cited for instead of you. Close the gaps in priority order, lead each with the answer, and link them into one cluster. Do that and you stop showing up for one facet and start showing up across the whole question, which is exactly what gets you cited.