What is query fan-out? The searches AI assistants run
In short
Fan-out searches are the web searches an AI assistant runs on its own before it writes an answer. One question becomes several searches, often longer and more specific than what the person typed. The pages those searches return are the pages the answer is built from.
Keats, from PressRoom. Updated Read as markdown
When you ask an assistant something that needs current information, it doesn't answer from memory. It writes its own searches, sends them, reads some of the results and answers from those. Those searches are called query fan-out, or fan-out searches. Google describes AI Overviews and AI Mode as using a "query fan-out" technique, issuing several related searches to build a response; ChatGPT does the same with its own searches.
They're often more specific than the question. Asked "which AEO tools track whether my brand shows up in ChatGPT", ChatGPT searched for "best AEO tools track brand visibility mentions in ChatGPT answers 2026 Profound Otterly Peec AI Ahrefs Brand Radar" (our sample). Note the year and the brand names: its search already leaned toward products it knew.
Why it matters: the brands in an answer are, mostly, the brands on the pages those searches return. If you want to know why an assistant names a rival, read its fan-out searches and the pages they found. That's also how you find which page to improve, and which Google AI Overview or roundup to get onto. Keats records the searches with every answer it reads, and shows them in each question's evidence.
Sources
- Google Search Central: AI features and your website, read November 12, 2026
- Keats: eight ChatGPT answers to buyers' questions about AEO tools (CSV), read November 12, 2026