Where a search engine returns documents for the user to evaluate, an answer engine evaluates the documents itself and returns a conclusion: a recommendation, a comparison, a definition, a shortlist. The user may never see the underlying pages, which is why being one of the sources the engine trusts has become a visibility goal in its own right.
How answer engines build an answer
- Model knowledge: facts absorbed during training.
- Retrieval: live web search and ranking of candidate sources, sometimes with query fan-out into multiple sub-searches.
- Synthesis: composing the answer, deciding which brands to name and which sources to cite.
- Presentation: citations, follow-up suggestions and, increasingly, transactional steps like shopping research.
Each engine weighs these layers differently and changes frequently, which is why visibility must be measured per engine and over time rather than assumed from a single test.