Learn
Query fan-out is the way an AI engine expands one question into several related sub-questions, answers each of them, and combines the results into a single response. Google documents the behavior for AI Overviews and AI Mode, and buyers do something similar by hand when they describe their situation instead of asking for the best tool in a category.
A question like "what should a five-person agency use to track client reporting" is not answered as one lookup. The engine breaks it into pieces: what agencies of that size need, which tools exist for client reporting, what they cost, how they compare. Each piece retrieves or recalls its own material, and the final answer is assembled from those. A brand is named or skipped at the level of the pieces, not at the level of the original question.
The practical consequence is that visibility is a map over many buyer questions, not a single "best X" prompt. A brand can be absent from the shortlist question and still be named when the buyer asks about a specific use case, or the reverse. Measuring one prompt and reading it as the verdict misses the structure the engine is actually using.
It also changes what to publish. If the engine answers sub-questions, the pages that help are the ones that answer a sub-question plainly: what the product is for, who it fits, what it costs, how it differs from the obvious alternative.
The audit derives the buyer's questions from the brand's own site and asks each one separately, so the result reads as a map: "recommended in 3 of 12 open questions", with the questions that named a competitor and the ones nobody answered clearly. The reasoning behind that design, with an example of a brand at 0 of 12, is in the blog post linked below.
Google describes it for AI Overviews and AI Mode. Other engines do not all document their internals, so we do not claim how each one works. The useful part for a brand is the same either way: buyers ask many specific questions, and the measurement should cover them.
Not from the engine. You can approximate them by listing the questions a buyer would ask about your category, which is what the audit does from your own site.
They rhyme but are not the same. Long-tail keywords are search queries with low volume. Fan-out sub-questions are generated by the engine from one request, whether or not anyone ever typed them.
Other definitions
The free audit reads your site, derives your buyer questions and measures them on three model APIs. No signup, no card.