An answer instead of a list of links
AI search finds web pages relevant to your question, reads them, and writes the answer out for you. Rather than returning a ranked list of links, it collects the relevant parts from several pages and presents them together with citations.
Perplexity, ChatGPT Search, and Gemini all offer this. It suits questions that need current information or a comparison across sources — what happened today, or how two products differ.
How it differs from ordinary search
A conventional search engine ranks pages against your keywords. You open the links, read, compare, and reach your own conclusion.
AI search does part of that middle step for you. You can ask in ordinary language and narrow down with follow-up questions. The trade-off is that which sources it used, and how it condensed them, is harder to see.
How it differs from Deep Research and RAG
AI search answers a single question from a relatively quick pass over the web. Deep Research spends much longer, covers more sources, and produces a long report.
RAG is the broader pattern of retrieving relevant material — not necessarily from the web, often from internal documents or a registered dataset — and feeding it into the answer. AI search can resemble RAG with the web as its corpus, though the internals differ by service.
Checking the answer
- Open the citations, not just the summary
- Separate the publication date from the date the event occurred
- Distinguish official announcements from third-party coverage
- Cross-check anything important against more than one source
- Confirm that the quoted passage means what the summary says it means
AI search speeds up the start of research. A visible citation does not guarantee that the cited page supports the claim. For decisions about health, law, money, or contracts, go to the primary source or a professional rather than stopping at the summary.