Query fan-out
Query fan-out is the way AI search engines work when the model splits one user question into several smaller queries and searches the web for each of them. It then builds a single, coherent answer from the results.
How it works
When a user asks a complex question, the model doesn’t search for the whole question at once. It first works out what information it needs, then runs separate searches for each aspect. Google AI Overviews and AI Mode use this approach, and so do ChatGPT and Perplexity when they search the web.
Why fan-out queries matter for your brand
An AI answer is built from pages found through the fan-out queries, not only from results for the question the user typed. If your page answers the sub-queries the model actually runs, it has a better chance of becoming a source, and your brand a better chance of being recommended. Knowing the fan-out queries in your industry helps you plan content: it shows which topics, comparisons and details are worth having on your site.
How to see fan-out queries
Some tools predict fan-out queries from the question; others record the queries the model actually ran for a real answer. Heuriso records fan-out queries from ChatGPT and Perplexity answers, and the action plan suggests which content to add or update for them.
Example
The question “Which invoicing software should a freelancer use?” can be split into queries such as “invoicing software for freelancers”, “free invoicing software comparison”, “invoicing software reviews” and “invoicing software bank integration”.