Query fan-out is how Google's AI Mode answers a question: it splits one search into many, runs them at once and writes one answer from what comes back.
Query fan-out is the technique Google’s AI Mode uses to answer a question. Instead of running your search once, it splits it into a set of smaller related searches, runs them all at the same time, and writes one answer from what comes back.
Google’s own guidance says both AI Mode and AI Overviews may use it. Robby Stein, Google’s VP of Product for Search, has described it as the system thinking up a list of related questions for you and then searching for each one, using Google Search as its own back-end tool.
That changes SEO. The search your page competes for is no longer only the one the person typed. It’s every sub-query Google writes on their behalf, and most of those never show up in a keyword tool.
How query fan-out works, step by step
- Google reads the question. A custom version of Gemini works out what the person is trying to do, how complex the question is, and whether it needs fanning out at all. Simple questions, like a capital city or a sum, skip the process.
- It writes its own searches. The model breaks the question into subtopics and turns each one into a search. “Things to do in Sheffield with kids” might become searches for rainy-day activities, family-friendly restaurants, free attractions and opening times.
- It runs them all at once. The searches go to Google’s normal web index, plus live sources such as the Shopping Graph, Maps and Google Finance.
- It reads passages, not whole pages. From the results it pulls the specific sections that answer each sub-question.
- It writes one answer with links. The model combines those passages into a single response and cites the pages it used.
If there are gaps after the first round, the system can run a second wave of searches. Deep Search, the heavier research mode inside AI Mode, can run dozens or even hundreds of background searches and take several minutes to finish.
Query fan-out vs traditional search
| Traditional search | AI Mode with query fan-out | |
|---|---|---|
| Searches per question | One | Many, written by Google |
| What gets judged | Whole pages | Passages within pages |
| What the person sees | A list of links | One written answer with cited sources |
| What keyword tools show | Most of the demand | Only the starting question |
What query fan-out means for your SEO
You compete on topics, not single keywords
A page built around one exact phrase can rank well in normal search and still be skipped in AI Mode, because Google’s sub-queries ask about things the page never covers. Pages that answer the main question and the obvious next questions get pulled into more of those searches.
Each section has to make sense on its own
Google lifts passages, so a section that only makes sense after reading the three above it is hard to use. Start each section with the answer, write headings that say exactly what’s underneath, and put the detail after.
Keyword volume understates real demand
Fan-out searches are written by Google, so they rarely appear in keyword tools with any search volume. A question showing 20 searches a month might feed hundreds of AI Mode answers. Plan content around the questions your customers actually ask, not only what the tools report.
How to optimise for query fan-out
- Map the follow-up questions. For each core topic, list what a buyer would ask next: cost, comparisons, risks, how long it takes, who it’s for. People Also Ask boxes, sales calls and support tickets are good places to find them.
- Cover them on the page, or link to the page that does. Build a pillar page for the topic and supporting articles for the deeper questions, all linked together.
- Write answer-first sections. Open each section with a one or two sentence answer, then explain.
- Use structure Google can read. Literal headings, short lists, comparison tables and schema markup make passages easier to pull out.
- Show who wrote it and why they know. Named authors, first-hand examples and cited sources help when Google checks several pages against each other. This is the E-E-A-T work that sits behind good content.
- Keep facts current. Prices, features and dates go stale quickly, and AI answers favour pages that are up to date.
- Keep ranking in normal search. Fan-out runs on Google’s standard index. If your pages don’t rank anywhere for the sub-queries, they won’t be pulled in.
How to tell if it’s working
Google doesn’t show fan-out queries anywhere, and AI Mode traffic is counted inside your normal web search data in Search Console. So you track it from the outside. Check your most important questions in AI Mode and AI Overviews each month, note which pages get cited, and watch your share of model across AI tools.
Rising impressions with flat clicks is often a sign your pages are being read inside AI answers rather than clicked. Our guide to zero-click search explains how to read that pattern.
If you want help working out which questions your customers are asking AI, and getting your pages into the answers, that’s what our AI SEO team does every day.
Common questions
Is query fan-out the same as AI Overviews?
No. AI Overviews is a feature: the summary at the top of a normal results page. Query fan-out is a technique Google may use inside AI Overviews and AI Mode to gather information for the answer.
Does query fan-out happen on every search?
No. Simple, factual searches usually skip it. It’s used when a question has several parts or intents, such as comparisons, plans and purchases.
Do other AI tools use query fan-out?
Google is the one that names it publicly, but ChatGPT search, Perplexity and Gemini’s deep research all break questions into several web searches in a similar way. The same content approach works across all of them.
Can I see which fan-out queries Google ran?
Not directly. Google doesn’t publish them or show them in Search Console. The closest you can get is testing questions in AI Mode and looking at which sources it cites.
