Retrieval: becoming a candidate source
Before a model can cite a page it has to find it. Generative engines pull candidates from a search index, a vector store, or a live browse step, and the filters at that stage are conventional: the page must be crawlable, indexed, fast enough to fetch, and topically close to the query. Content locked behind client-side rendering, blocked by robots rules, or buried on a page about six other subjects rarely reaches the shortlist.

