What Are Google AI Overviews and AI Mode? How to Become a Source in AI Summaries

What are Google AI Overviews and AI Mode, and how do they differ from classic results?

AI Overviews (AI Bakışı in the Turkish interface) is the AI response Google shows at the top of the results page for some searches, summarising the essentials of a topic with links to web sources. AI Mode (AI Modu) is a conversational search experience that moves forward through follow-up questions, built for questions that call for more exploration, reasoning or comparison.

In classic results, the user picks from a list of links. With these two features, they first read an answer written by AI and, if they want more detail, go on to the links next to that answer. Google describes AI Overviews as a feature that helps people quickly grasp the gist of a complex topic and acts as a jumping-off point for further exploration. We explain how classic results come together in our what is SEO guide.

In Türkiye, both features have been rolling out gradually in Turkish since February 18, 2026, on the search results page and in the Google app on Android and iOS. According to Google Türkiye’s announcement, AI Mode is powered by the Gemini 3 model family, while ads continue to appear in dedicated slots above and below the AI Overviews panel. We cover the difference between paid and organic visibility in our article on SEO vs SEM.

Users can’t turn AI Overviews off, but they can use the Web filter on the results page to see results made up of links only. Google also states plainly that AI responses can contain mistakes. Measuring how your brand is described in these responses with our AI visibility test is the first step towards spotting incomplete or inaccurate descriptions.

Which searches trigger AI Overviews?

Google doesn’t show AI summaries on every search, only on searches where its systems decide that generative AI will be especially helpful for that query. In practice these are mostly long, question-style queries that require combining information from more than one source.

One independent dataset that shows this comes from Pew Research Center. In a study published in July 2025, which looked at the Google searches of 900 US adults in March 2025, about one in five searches (18%) produced a summary. A summary appeared on 60% of searches starting with question words such as who, what, when or why, and on 53% of searches of ten words or more, while the rate fell to 8% for searches of one or two words.

• Summary more likely: question-style searches such as “how”, “why” or “which”; long queries with several conditions; explanations of concepts and comparisons.
• Summary less likely: short queries of one or two words.
• Where AI Mode fits: according to Google, AI Mode is designed for asking detailed questions in one go that would previously have taken several separate searches, and for making complex comparisons.

Whether a summary appears can vary by user, location and time, and Search Console doesn’t report this distinction either. That is why you need to check your target query set at regular intervals and record which queries show a summary and who gets cited. We explain how to work out mention rate from these records in our article on SEO metrics.

How does Google choose sources for AI Overviews?

Google chooses the sources for AI summaries from its own search index: with a technique called query fan-out, which splits the question into subtopics, it can run multiple related searches, and it shows as links the pages that support the answer, drawn from pages that are indexed and eligible to be shown with a snippet.

The practical consequence of this technique is significant. According to Google, query fan-out makes it possible to show a broader and more diverse set of links than a classic web search. In other words, a page that doesn’t rank first for the main query but answers one subquestion most clearly can still be a source in the summary. In Pew’s March 2025 data, 88% of summaries cited three or more sources.

• Cover the subquestions. A page, or a cluster of interlinked pages, that also answers the side questions a user might ask alongside the main question shows up in more branches of the fan-out.
• Answer at passage level. A clear answer at the start of each section creates a passage that can be used for that subquestion; we cover the rules of citable content in detail in our what is GEO guide.
• Protect snippet eligibility. nosnippet or a very low max-snippet value limits how the page can be used in summaries.
• Make trust signals visible. Google says its systems aim to surface content that demonstrates experience, expertise, authoritativeness and trustworthiness (E-E-A-T); author, date, sources and company information are how these signals show up on the page.

Google doesn’t publish its detailed selection criteria, but it does state that the fundamental practices of SEO apply to AI features as well. So the groundwork for becoming a source in summaries is the same as for on-page SEO: a page that matches search intent, a clear heading structure and strong content.

Do you need special optimisation to appear in AI Overviews?

No. According to Google, there are no additional requirements and no special optimisation needed to appear in AI Overviews or AI Mode. The technical prerequisite is simply that the page is indexed and eligible to be shown in Search with a snippet; the rest comes down to the fundamentals of classic SEO.

Google also sets out, in its developer documentation, the fundamental practices that apply to these features.

• Allowing crawling in robots.txt and in your CDN or hosting infrastructure.
• Making content easy to find through internal links.
• Providing users with a good page experience.
• Making important content available as text.
• Supporting text with high-quality images and videos where appropriate.
• Making sure structured data matches the visible text on the page.
• Keeping Merchant Center and Business Profile information up to date.

The same document also says there is no need to create new machine-readable files, AI-specific text files or special markup to appear in these features, so files such as llms.txt are not a prerequisite for Google’s summaries. Schema isn’t a ticket to rich results or summaries either; it is a layer that makes content easier to understand, and we explain how to set it up in our article on structured data.

Showing different text only to AI systems, on the other hand, is a risk rather than a solution: serving different content to users and to search engines is defined as cloaking in Google’s spam policies. That is why we don’t handle work aimed at AI summaries by producing a separate version; we run it in the same plan as the technical and content decisions in our GEO and SEO consultancy.

How do AI Overviews affect clicks and organic traffic?

AI summaries can reduce clicks on informational queries in particular, because they make it easier for users to get the gist of the answer on the results page; the impact, however, varies from site to site and by query type. The data available shows both sides together.

Independent measurement: In a study Pew Research Center published in July 2025, based on the browsing data of 900 US adults in March 2025, users followed a link to a traditional result in 8% of visits to pages with a summary, compared with 15% on pages without one. Following a link inside the summary itself happened in just 1% of visits, and users were also more likely to end their browsing session after a page with a summary (26% versus 16%).

Google’s account: Liz Reid, head of Google Search, wrote in August 2025 that the total volume of organic clicks from Google to websites had stayed relatively stable year over year, while quality clicks in which users don’t head straight back had risen slightly; the same post also acknowledges that traffic fell for some sites and grew for others. Google’s developer documentation likewise states that clicks from pages with summaries are higher quality, meaning users are more likely to spend more time on the site.

Reading the impact on your own site means looking at indirect signs, because Search Console shows clicks from these features within the Web search type, without a separate breakdown. Informational queries whose click-through rate falls while impressions hold steady are the group most likely to have been affected by summaries. We cover the steps for diagnosing a drop in our article on organic traffic, and the Search Console side in our Google Search Console guide.

What can you do about a drop in clicks from AI summaries?

Summaries can’t be switched off, and pulling a page out of them usually lowers its visibility in classic results as well. The more productive approach is to rebuild your content and measurement plan around the new results page.

• Invest in content that needs a visit. Calculators, current pricing or stock information, comparison tables, original research and hands-on examples can’t be fully replaced by a summary.
• Track commercial and local queries separately. On queries with purchase or contact intent, users usually have to go through to a website; give these pages priority.
• Aim to be a source. Being shown as a link in the summary puts your brand in front of the user even when it doesn’t lead to a visit; measure this with mention rate.
• Tie success to business outcomes. If key events such as form submissions, calls and sales hold steady while sessions fall, the loss is in low-intent visits.

How do you limit a page’s use in AI Overviews and AI Mode?

A page’s use in AI Overviews and AI Mode is limited with Search’s preview controls: nosnippet, data-nosnippet, max-snippet and noindex. According to Google, the Google-Extended product token in robots.txt does not affect inclusion in Search, so it is not a control for these features.

When you decide, keep in mind that these controls don’t work separately for AI summaries; they work together with classic search previews. Blocking the whole page with nosnippet removes the description in classic results along with the summaries and can lower your click-through rate. In practice, setting aside only specific sections you don’t want quoted, such as the preview of members-only content, with data-nosnippet is a more balanced approach. Access for AI bots other than Google’s is managed separately in robots.txt; we explain this distinction in our AI crawlers guide.

To see which of your pages are cited in summaries today and which competitors are being recommended, a free GEO analysis is a good place to start. If you want to turn this work into an ongoing plan, the GEO items in our SEO packages bring it into the monthly plan.