What are AI Overviews? Definition, history and how they work

What are AI Overviews?

AI Overviews are a short synthesised answer at the top of the results page, with source links shown next to or inside the text. Google generates them only for queries where it believes a summary helps more than a list of links, typically longer informational questions, comparisons and how-to searches. The feature lives inside regular Google Search, so users do not opt in or switch tabs. The conversational sibling, Google AI Mode, is a separate tab that replaces the results list with a chat-style answer.

How do AI Overviews work?

AI Overviews work by retrieving pages from Google’s index and letting a custom Gemini model write a summary grounded in those pages. Google Search Central documentation describes the process in a few stages:

1. Interpret the query and decide whether an AI response adds value.
2. Run related searches across subtopics and data sources, a technique Google calls query fan-out.
3. Select supporting pages from the index that are eligible to show a snippet.
4. Generate the summary with Google Gemini and attach links to the pages it used.

Since May 2025 the US version runs on a custom version of Gemini 2.5, according to Google’s I/O announcement.

AI Overviews timeline

AI Overviews grew from a US launch to a global default feature in about 12 months.

How do AI Overviews affect SEO and traffic?

AI Overviews reduce clicks on classic organic results for queries where a summary appears. Pew Research Center tracked the browsing of 900 US adults in March 2025 and found that users clicked a traditional result on 8% of visits with an AI summary, compared with 15% without one. Only about 1% of those visits included a click on a link inside the summary itself, and 26% ended the browsing session right after the results page. Google presents a different angle: in its May 2025 update the company said AI Overviews drive over 10% more usage of Google for the query types that show them in the US and India.

For marketers this means two things. First, organic click-through rate (CTR) on informational keywords drops even when rankings stay stable, so I recommend tracking impressions and branded demand next to clicks (see my guide to measuring SEO results). Second, being cited inside the overview becomes a visibility goal of its own, which is the core of Generative Engine Optimization (GEO).

How to appear in Google AI Overviews?

A page appears in AI Overviews when it is indexed, eligible for a snippet and answers a sub-question better than competing pages. Google states that no special markup, AI text file or extra technical requirement exists. In my projects these practices raise the citation rate:

• Answer the question in the first sentence under each heading.
• Keep key facts in HTML text, not only in images or scripts.
• Cover related sub-questions, because fan-out searches look for them.
• Keep structured data consistent with the visible content.
• Allow Googlebot in robots.txt and in CDN firewall rules.

Traffic from AI Overviews appears in Google Search Console inside the Performance report under the “Web” search type, without a separate filter. Publishers who want to limit how much text Google uses can apply nosnippet, data-nosnippet or max-snippet.

AI Overviews vs AI Mode