AI Overviews and Query Fan-Out, Explained
What an AI Overview actually is
When Google shows an AI Overview above the normal search results, it’s not writing an answer from scratch out of what it already “knows.” It’s running a version of the same search process you’d expect from a normal query, then summarizing what it finds. Google has described the specific technique behind this as query fan-out: instead of running your one search, the system breaks it into a handful of related sub-questions, runs each of those searches separately, and stitches the results together into the answer you see. ChatGPT search, Perplexity, and Microsoft Copilot all work on a similar principle, even if the details differ.
That matters for how you think about “ranking.” You’re not trying to be the single best answer to one query anymore. You’re trying to be a good answer to several sub-questions the system generates on its own, sometimes for a query you never directly targeted.
What that changes about SEO, and what it doesn’t
The foundation doesn’t change. A page still needs to be crawlable, indexable, fast, and actually about the topic it’s trying to rank for. None of that goes away just because the results page looks different. If your technical SEO is broken, you were never in the running for an AI Overview citation in the first place.
What does change is how much a single page needs to stand on its own. If an AI system fans a query like “best digital marketing consultant for ecommerce” out into sub-questions like “what does a marketing consultant charge,” “do consultants manage Google Merchant Center feeds,” or “what’s the difference between an agency and a consultant,” a page that only vaguely gestures at all of those without directly answering any of them is easy to skip over in favor of a page (yours or a competitor’s) that answers one of those sub-questions clearly and specifically.
Practically, the things that tend to help are:
- Direct, specific answers to specific questions. Not because “answering questions” is some new AEO trick, but because that’s literally what a sub-query in a fan-out is: a question. FAQ sections with real, specific answers help here for exactly that reason, not because FAQ schema is magic.
- Structured data that makes it unambiguous what a page is about (Service, FAQPage, Article, etc.). It doesn’t guarantee a citation, but it removes the guesswork for a system trying to match your page to a sub-query.
- Genuine topical depth, since a fan-out can surface a page for a sub-question you didn’t explicitly write a section for, as long as the surrounding content makes clear you’re a credible source on the general topic.