What are Google AI Overviews?

What is AI Overviews and where does it appear?

AI Overviews is the AI summary that appears above a search query and answers it directly. Beneath or beside the summary sit links to the pages the answer was built from. What the user sees is not ten blue links but an answer first and its sources second — which fundamentally changes how attention is distributed on a results page.

The summary does not appear on every query. It shows up far more often on informational questions, on anything asking for an explanation, and on anything requiring several sources to be combined; on purely navigational queries — typing a brand name to reach its site — it usually does not appear at all. That distinction is the most practical way to predict which of your pages will be affected.

Where does AI Overviews pick its sources from?

There is no separate index. AI Overviews is fed by pages in Google's classic search index, so a page that is not in the index cannot appear in the summary. That single sentence clears away the most common misunderstanding on this subject: there is no separate channel, no separate submission and no separate technical requirement for entering the AI summary.

Google writes that it uses a technique it calls query fan-out when producing the answer: splitting one question into subtopics and data sources and running several searches. The practical consequence is this — a page that does not rank first for the main query can still be a source in the summary if it gives the best answer to one of the subtopics. Classic ranking and presence in the summary are not the same thing.

Google's own wording is unambiguous: 'There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.' That sentence is the shortest possible answer to any proposal selling a separate technical package for AI visibility.

Does blocking Google-Extended remove you from AI Overviews?

No — and not knowing that distinction gets expensive. Google-Extended is the token used to limit AI training and grounding in Google's other systems, the Gemini side. AI Overviews, by contrast, is built into Search, and the crawl control there is Googlebot and robots.txt.

The consequence: the only way to leave AI Overviews entirely is to leave Google search. Google says so directly — because AI is built into Search and integral to how Search functions, the control site owners have is the robots.txt directives written for Googlebot. There is no button reading "keep me in search but out of the summary".

How do I limit what the summary shows?

Google documents four controls for limiting the information shown from a page: nosnippet, data-nosnippet, max-snippet and noindex. They are not specific to the AI summary — they govern search snippets as a whole — and they are the only documented route.

<!-- Keep the whole page in search, but stop the price line
     from being used as a snippet: -->
<p>An identity project takes about six weeks.</p>

<p data-nosnippet>
  Starting budget for 2026: 180,000 TRY. This figure may change
  during promotional periods.
</p>

<!-- Cap the snippet length for the whole page: -->
<meta name="robots" content="max-snippet:160">

<!-- Show no snippet at all: -->
<meta name="robots" content="nosnippet">

How do I measure AI Overviews in Search Console?

There is no separate filter, and knowing that saves hours. Data from AI features is counted inside the Web search type in the Performance report, together with overall search traffic. So the question "how many clicks came from AI Overviews" has no direct answer in Search Console.

What can be done is an indirect reading, and it works: isolate the query groups where impressions rise while the click rate falls. That pattern is the strongest available indicator that a summary has appeared above the query. Doing it query by query gives a far cleaner result than doing it page by page, because the summary is triggered by the query, not by the page.

Is AI Overviews cutting traffic, and what should you do?

On informational queries the click rate can fall, because the user gets the answer inside the summary. On commercially intended queries — price, purchase, comparison, near me — users still go to the source, because the decision itself is made on the site. That distinction dictates how the response should be built.

• Do not withdraw the content. A brand named in the summary gains a marked position over a competitor that is absent; choosing invisibility enlarges the loss.
• Rewrite informational pages to be sources: a clean answer passage, question-shaped headings, sourced numbers.
• Protect and strengthen commercial pages; the summary does not satisfy the intent there, and the user comes to the site to decide.
• Do not chase both a definition and a purchase intent on one page; separating them wins on the queries that trigger a summary and on the ones that do not.
• Build measurement query by query. A page-level view hides the fact that summaries are triggered by queries, and leads you to blame the wrong page.