Listen: Query Fanout
The step where an AI system splits one prompt into several single-intent sub-queries, each retrieving its own sources; a page can be grounded for one angle of a question and absent for another.
Transcript
When you ask an AI a complex question, it rarely searches for the answer in just one go. Instead, it uses a process called query fanout. This is where the system takes your single prompt and splits it into several smaller, single-intent sub-queries.
Each of these sub-queries then goes off to find its own specific set of sources. Because the search is split up this way, a single web page might be highly relevant and used for one specific angle of your question, but completely ignored for another.
Ultimately, this targeted search process is what feeds and shapes the final grounding snippets the AI presents to you, ensuring each part of your complex question gets a precise, well-sourced answer.