Answer engine

A search engine returns a ranked list; an answer engine returns a written answer. Underneath, most answer engines still search. They take the question, run one or more web searches, read the top results and write a reply with the sources in context, often citing them inline. That last step is why the term matters to a store: the reply is only as good as the pages the engine read, and those pages are chosen by search relevance, recency, and how directly they answer. Product pages, articles, reviews and forums can all be sources. Match the source to the question rather than assuming one page type always wins. Each engine reads with its own crawler and its own rules. ChatGPT, Perplexity, Claude and Google each publish which user agents fetch pages for training, for their search index and on a user's behalf at answer time, and each respects robots.txt in its own way. Being present in an answer engine therefore has two parts: letting the right crawlers in, and having pages worth quoting when they arrive. The AI crawler guide distinguishes training, search and user-triggered access before you change a rule.