Google AI Overviews and SEO: What Changes for Organic Visibility

An AI Overview is a short answer produced by a Gemini model customised for Search. It sits at the top of the results page and includes links to the web pages it used. Users can expand it, follow the cited links or scroll past to the regular results.

Google has said that AI Overviews appear when its systems judge that a generated summary adds value beyond the standard results, which tends to mean multi-part or exploratory questions. Simple navigational searches, such as a brand name, usually do not trigger one. Google also offers AI Mode, a separate conversational experience, which uses similar technology but works as its own tab.

For some queries, yes. When the overview fully answers a simple question, fewer people need to click anything. Publishers that relied on short definitional content have reported drops in click-through for those terms. Google has argued that clicks from AI Overviews are higher quality, meaning visitors stay longer, but individual sites see different results.

The effect is uneven across a site. A typical pattern looks like this:

• Definition and “what is” pages lose the most clicks, because the summary covers them.
• Comparison, how-to and pricing pages hold up better, because users want detail and specifics.
• Branded and transactional queries are rarely affected.
• Local queries are shaped more by the map pack than by the overview.

So the honest answer is that total traffic may fall on some page types while staying steady on others. Measuring by page type tells you far more than a sitewide average.

Google has not published a separate ranking system for AI Overviews. Its guidance says the same fundamentals apply: pages must be indexed and eligible to show a snippet, and the content should be helpful, reliable and written for people. There is no special markup that guarantees inclusion.

In practice, cited pages tend to share some traits. They answer the question directly near the top, they break complex topics into clear sections, and they come from sites that already rank for related terms. Overviews often cite pages that are not in the top three organic results, which gives well-structured content on smaller sites a real chance.

The core of SEO stays, but the emphasis moves. Thin pages written only to capture a keyword are now competing with a free summary. Pages with original detail, first-hand experience or data the model cannot invent become more valuable.

Practical changes most sites can make:

• Open each article with a two- to three-sentence answer to the main question.
• Use question-style subheadings that mirror how people search.
• Add specifics: steps, settings, example numbers you can stand behind, screenshots.
• Keep author information and update dates visible and accurate.
• Make sure the snippet controls are not blocking you; a “nosnippet” directive also removes content from AI Overviews.
• Strengthen internal links so related pages are discovered and understood as a group.

This is the awkward part. Search Console counts impressions and clicks from AI Overviews inside the normal web search totals, without a separate filter at the time of writing. A link inside an overview reports the position of the overview as a whole.

That means you need to combine data sources. Track impressions and click-through rate by query group in Search Console, and watch for queries where impressions rise while clicks fall. Several rank-tracking tools now record whether an overview appeared for a keyword and whether your domain was cited. Manual spot checks in a clean browser session help too, though results vary by location and user.

Signals worth reporting monthly

• Click-through rate by page type, compared to the same period last year.
• Number of tracked queries showing an overview.
• Share of those where your domain is cited.
• Conversions from organic search, which matter more than raw clicks.

Generated answers do not remove the need for a healthy site. If Googlebot cannot crawl a page, render its main content or understand which version is canonical, that page cannot be used as a source. Technical problems that once cost a few positions can now cost both the ranking and the citation.

Run through the basics on the pages you most want cited. Confirm they are indexed in the URL Inspection tool, that the main text is present in the rendered HTML rather than loaded only after user interaction, and that structured data matches what is visible on the page. Check page speed too: Core Web Vitals do not decide citations on their own, but slow, unstable pages tend to perform worse across search features. Finally, make sure each important page has a clear, descriptive title and a single main topic, which helps every system that summarises it.