Lesson 2: How Google AI Overviews and AI Mode Pick Sources

Why this level matters

In Level 1 you separated two ways an assistant can know about you: what it learned during training, and what it fetches live. Google AI Overviews and AI Mode belong to the live-retrieval group. They build answers from the Google Search index, the same one that drives the ordinary blue links. So if a page is not in that index, it cannot be a source.

This lesson covers what Google's own documentation says (pages read on 2026-10-20). It does not cover industry theories about ranking. Those come in Level 4.

What Google documents about source selection

Google's Search Central page "AI features and your website" (last updated 10 December 2025) makes four points:

1. The requirements are the same as for normal Search. To be shown as a supporting link, a page must be indexed and eligible to appear in Google Search with a snippet. Google says there are "no additional requirements" and no special optimisations needed.
2. No new files or markup are needed. Google states you do not need new machine-readable files, AI text files or special schema.org structured data to appear in these features.
3. Query fan-out. AI Overviews and AI Mode may run several related searches across subtopics and data sources to build one answer. Google says this can show a wider and more diverse set of links than a classic results page. In practice, a page that ranks for a narrow subtopic can be cited for a broad question it never ranked for directly.
4. Reporting. Clicks and impressions from AI features are counted in the Search Console Performance report under the "Web" search type. They are not split out separately there.

An analogy: picture AI Overviews and AI Mode as a researcher who works only in Google's library. Being on the shelf is indexing. Being allowed to quote you is snippet eligibility. Fan-out means the researcher goes to several shelves for one question.

Which controls affect AI Overviews and AI Mode

These controls do apply, because they are ordinary Search controls:

• Googlebot access in robots.txt. If Googlebot is blocked, the page cannot be crawled for Search, and so it cannot be crawled for these features either.
• noindex removes the page from the index completely: no blue link and no AI citation.
• nosnippet stops any text from the page being shown. Google's documentation names it as a way to limit what appears in AI features.
• data-nosnippet hides specific parts of a page.
• max-snippet:[number] caps how much text can be shown.

This control does not remove you from AI Overviews or AI Mode:

• Google-Extended. This robots.txt token controls whether Google may use your content to train future Gemini models and to ground answers in the Gemini Apps and Vertex AI. Google's crawler documentation (updated 15 October 2026) says Google-Extended "does not impact a site's inclusion in Google Search nor is it used as a ranking signal." AI Overviews and AI Mode are part of Search, so blocking Google-Extended leaves you eligible for them.

Safety warning: every control that removes you from AI Overviews (Googlebot block, noindex, nosnippet) also weakens or removes your normal Search listing. No documented switch removes you from AI features while keeping full blue-link visibility. Test any change on one page before applying it across the whole site.

Task 1: Check your controls

1. Open yoursite.com/robots.txt. Note any rules for Googlebot and for Google-Extended.
2. On your 5 most important pages, view the source and search for noindex, nosnippet, max-snippet and data-nosnippet.
3. In Search Console, run URL Inspection on each page and confirm it shows "URL is on Google".
4. Write down what you found for each page, and whether you put each restriction there on purpose.

Task 2: Run a citation audit for 10 priority queries

1. List 10 queries where you are losing clicks, for example "best CRM for small accountants" or "how to descale a [your product]".
2. Open a private browser window, signed out, set to your target country.
3. Search each query. Note whether an AI Overview appears and which domains it links to.
4. Run the same query in AI Mode and record the cited sources.
5. Record these columns: query, date checked, country, AI Overview shown (Y/N), your pages cited, competitor pages cited.
6. Repeat the audit monthly. Answers change, so a single check is only a snapshot.

Task 3: Record AI Mode availability in your markets

In its blog post of 7 October 2025, Google announced AI Mode in more than 35 new languages and over 40 new countries and territories, for a total of more than 200 countries and territories. Availability still changes by country, language and feature.

1. Open Google's Search Help page for AI Mode and the latest AI Mode posts on blog.google.
2. For each market you sell into, write down: whether AI Mode is available, which languages it supports, and any feature labelled as an experiment, Labs feature or preview.
3. Note the date you read each page.

How to measure it

• Count your citations across the 10 queries each month. Track it as a share, for example 3 of 10 queries cite you.
• In Search Console, compare clicks and CTR for those queries over time in the "Web" search type. Remember that AI-feature traffic is included in these figures and is not reported on its own.

Key takeaways

• AI Overviews and AI Mode cite pages from the normal Google Search index. To be eligible, a page must be indexed and allowed to show a snippet.
• Google says no special schema, AI text file or extra optimisation is needed.
• Query fan-out means one answer can draw on pages that rank for related subtopics.
• noindex, nosnippet, data-nosnippet, max-snippet and Googlebot access affect these features. Google-Extended does not.
• Always record the date you checked a citation or an availability page.