Query fan-out
Query fan-out is how an AI search or shopping system turns one question into many smaller searches. It splits a single prompt into several parallel sub-queries, retrieves results for each, and combines them into one answer.
How it works
When a shopper asks for "a gift for a friend who just started running," the engine rarely searches those exact words. It fans the question out into narrower searches: beginner running shoes, running socks, a hydration belt, a sensible price band. It retrieves candidates for each of those hidden searches, then writes one answer from what it found.
A commerce example
You are not competing only for the shopper's literal question. You are competing for the dozen reformulated searches the model invents, and many of them carry hard filters: a price ceiling, a size, a color, in stock now. Your product can be a perfect match for the shopper and still lose because it was not eligible for the specific sub-query the model ran.
What this means for your store
This is the layer where visibility is actually decided, and most store owners never see it. The practical response is the same as the rest of agentic commerce optimization: be retrievable and complete for the real sub-queries in your category, not just your one obvious keyword. Complete attributes and accurate availability are what let you win filtered sub-queries.