AI Overviews

AI Overviews can satisfy a query without the user clicking any result — a "zero-click" answer. Being cited in one depends on the same signals that drive GEO: a clear, direct answer, credible corroboration, and content a model can parse and quote confidently.

AI Overviews are generated by a large language model reading the web. To understand how that works end to end, see What Are Large Language Models (LLMs)?

How it works

For queries where Google judges a summary useful, it runs a language model over results it has retrieved and writes a short answer at the top of the page, with links to the pages it drew on. Which pages are cited depends on retrieval and on how clearly each page answers the specific question, not only on classic ranking position.

Overviews appear most for informational and comparison questions, and less for simple navigational or transactional searches.

Example

A search like "how long does it take to build a custom website" may show an overview that states a typical range and cites two or three pages. A page that answers in its first paragraph, with the range stated plainly, is far easier for the model to quote than one that buries the answer under an introduction.

Common mistakes

• Measuring success only by clicks, when being cited is now part of visibility.
• Writing long introductions before the answer the searcher came for.
• Blocking search crawlers from content and then expecting to be cited.