Google AI Overviews & AI Mode: How to Get Visible

To appear in Google AI Overviews or AI Mode, a page first has to be indexed and eligible to show a snippet in ordinary Google Search. Google requires no separate submission, special file or AI-specific markup. Past that point, whether your page becomes a supporting link depends on how well it answers the many sub-questions Google's models generate behind the scenes, and on the usual quality signals. Nothing guarantees inclusion, so treat it as something you measure, not something you can switch on.

Key takeaways

Eligibility is plain SEO. Crawlable, indexed, snippet-eligible; nothing extra.
Query fan-out changes the game. Google runs related sub-searches, so a page can be cited for a sub-question even when it doesn't rank for the main query.
Answer-first, evidence-rich passages win. Self-contained paragraphs that define, compare and quantify are easier to lift and ground.
Controls cost something. nosnippet and similar rules limit AI features and classic snippets alike.
Measurement is blurry. Search Console counts AI feature traffic inside the normal "Web" search type, so you need your own sampling to see where you actually appear.

What are AI Overviews and AI Mode?

AI Overviews are generated summaries shown at the top of a normal results page for some queries. Google launched them in the US in May 2024 as the successor to its Search Generative Experience. In July 2025 Sundar Pichai said they had more than 2 billion monthly users. AI Mode is a separate, conversational search interface built on a custom version of Gemini. It arrived as a Labs experiment in March 2025, opened to all US users at Google I/O in May 2025, and in October 2025 Google announced an expansion to more than 200 countries and territories, with over 35 new languages. Availability still differs by market and language.

	AI Overviews	AI Mode
Where it appears	Above classic results, for some queries	A separate tab or entry point
Interaction	One summary, with optional follow-ups	Multi-turn conversation
Typical queries	Informational, "how/what/why"	Complex, comparative, exploratory
Retrieval	Search index plus query fan-out	Heavier query fan-out, more sources per answer
Links	Supporting links inside and beside the summary	Inline links and a source panel

Both fall under generative engine optimization, but unlike ChatGPT or Perplexity they draw directly on Google's index and ranking systems, so the differences between GEO and SEO are smaller here than anywhere else.

How does Google choose the sources in an AI summary?

Google's documentation describes a technique called query fan-out: to build a response, AI Overviews and AI Mode "may issue multiple related searches across subtopics and data sources." While the answer is being generated, the system finds more supporting pages, so the set of links can be wider and more varied than a classic top ten.

For example, "best CRM for a 10-person agency" might fan out into "CRM pricing for small teams" or "agency CRM integrations." A page that clearly answers one sub-question can be cited even from page two for the head term, which is why classic rank and AI inclusion correlate but don't match.

Google hasn't published a ranking formula for supporting links. What is safe to say: candidates come only from indexed, snippet-eligible pages; the model grounds its answer in retrieved passages, so explicit statements beat implied ones; and Google's core quality systems still apply. For more on how generative systems pick evidence, see which sources AI cites.

Technical eligibility: a step-by-step checklist

According to Google Search Central, there are "no additional technical requirements" for AI features. In practice, that means getting the fundamentals right:

Allow crawling. Check that robots.txt, your CDN and any bot protection don't block Googlebot.
Get indexed. Use URL Inspection in Search Console. Pages blocked by noindex or canonicalized elsewhere can't be cited.
Keep snippets allowed. A page with nosnippet can't serve as a supporting link.
Put key content in text. Facts hidden in images, PDFs-only or client-side widgets are harder to use. Add images and video as support, not as a replacement.
Link internally. Make important pages reachable within a few clicks.
Keep structured data truthful. Markup must match visible text. Google states no special schema is needed for AI features.
Update Merchant Center and Business Profile if you sell products or serve a local area.

Google also says no new machine-readable files or AI text files are needed for these features. A file like llms.txt may serve other tools, but it's not a ranking input for AI Overviews.

What content gets cited in AI Overviews and AI Mode?

Once a page is eligible, content decides. Four patterns follow from how grounding works:

Answer first, then elaborate

Open each section with a one- or two-sentence direct answer. A passage that names its subject and doesn't rely on "as mentioned above" can be extracted cleanly.

Cover the sub-questions, not just the head term

Map the questions a user would ask next: costs, comparisons, edge cases, steps. One deep page with clear headings can cover several fan-out branches.

Add verifiable evidence

In the GEO paper by Aggarwal et al. (KDD 2024), adding statistics, quotations and citations to sources improved visibility in generative engine responses by up to about 40% in their benchmark. That was a benchmark, not Google itself, but the mechanism fits. A practical guide is in how to write content AI quotes.

Show experience and accountability

Author names, dates, first-hand data and clear sourcing match Google's people-first content guidance. They make your page a safer citation.

Hypothetical example: a Berlin accounting firm publishes a dated table of VAT thresholds for freelancers, linked to the official source. For "do I need a VAT number as a freelancer in Germany," it can be cited for the threshold sub-question even if a large publisher outranks it.

Controls: limiting how your content appears

Google's controls work through normal Search mechanisms:

nosnippet stops snippets and use as a direct input to AI Overviews and AI Mode.
data-nosnippet excludes specific parts of a page.
max-snippet limits snippet length.
noindex removes the page from Search entirely.

The trade-off is real: these settings also shrink your classic snippets. Google-Extended, the token that limits use of content for training and grounding in some other Google AI systems, does not remove you from AI features in Search. Those follow Googlebot and the preview controls above. After a change, verify with URL Inspection and allow time for recrawling.

How do you measure visibility in Google's AI summaries?

Search Console counts clicks and impressions from AI Overviews and AI Mode inside the regular "Web" search type, with no separate filter. Google says clicks from pages with AI Overviews are "higher quality," but it doesn't publish data to support that.

Independent data suggests caution about traffic. In a Pew Research Center analysis of March 2025 browsing data from 900 US adults, users clicked a traditional result in 8% of visits when an AI summary appeared, against 15% without one. They clicked a link inside the summary in only 1% of visits.

So AI visibility needs its own measurement: a fixed set of real questions, run repeatedly per market and language, with the cited sources stored. Because AI answers vary from run to run, a single screenshot proves little. A presence rate over a sample does. Our methodology describes one way to do this with honest denominators, and the steps are covered in how to measure AI visibility. GeoDog applies the same approach to ChatGPT, Gemini, Claude and Perplexity, which helps you see whether your Google gains carry over to other assistants.

FAQ

Do I need special schema or an AI file to appear in AI Overviews?

No. Google states that no special schema.org markup or AI text files are required. Indexing and snippet eligibility are enough to be considered.

Does ranking first guarantee a place in the AI summary?

No. Classic rank helps, but query fan-out means sources can come from sub-queries, and the model picks passages it can ground. Lower-ranked pages are sometimes cited, and top-ranked ones sometimes aren't.

Can I stay in Google Search but opt out of AI Overviews?

Not cleanly. The available controls (nosnippet, data-nosnippet, max-snippet) also affect classic snippets, and Google-Extended doesn't apply to AI features in Search.

Google AI Overviews and AI Mode: how to get visible in AI summaries