Query Fan-Out When a user enters a complex prompt, the answer engine translates it into multiple clean search queries. One messy human question becomes several focused searches, each returning its own list of potential sources.
The Cover (Billboard) The only information an answer engine sees before deciding to click: URL, title, snippet, and sometimes freshness date. This is what determines whether your page gets opened, regardless of how good the content inside is.
Decision-Ready Content Content structured so the engine can lift it directly into an answer. Utility-first, answer-shaped blocks rather than narrative or single-angle pieces.
The Core Rule The model can only answer from what it fetched. Not what exists. Not what is true. What it actually pulled into context.
Three Constraints of Retrieval
Diagnostic Framework: Why Am I Not Showing Up?
Gate 1: Fetchable Can the engine reach and ingest your page?
Gate 2: Chosen Does your cover look like the answer?
Gate 3: Extractable Can the engine lift a clean chunk from your page?
Extraction Principle Retrieval is text-first. If your answer lives only inside an image, chart, or video without surrounding text, most systems cannot extract it.
