AI Overviews

How it works

AI Overviews are Google's generated answers rendered above the traditional result list. The system takes the query, expands it into several related sub-queries (see query fan-out), retrieves passages for each from the search index, and synthesizes a single answer with inline links to the sources it drew from. The citation set is chosen per sub-query, which is why the pages cited frequently are not the pages ranking first for the original query.

When it matters

It matters because it changes what a top ranking is worth. An overview that answers the question resolves a share of queries without a click, so the value shifts from position to inclusion in the citation set. For any informational query in your category, the practical question stops being "do we rank" and becomes "are we one of the sources the answer was assembled from". Measuring that is a different exercise from rank tracking. See How AI Search Works and AI Visibility Audit.

Common misconceptions

• "Ranking first gets you cited." Retrieval runs per sub-query against the index, so pages well outside the top organic results are routinely cited while the number one result is not.
• "It only affects informational queries." Commercial comparison queries, which is where most B2B software research starts, are heavily covered.
• "Blocking the crawler protects your traffic." It removes you from the citation set while the overview still gets written from someone else's pages.