Guide

Google-Extended Does Not Affect AI Overviews: Three Controls That Do

Summary

No. Blocking Google-Extended changes whether Google may use a page to train and ground its Gemini models. It leaves Search, AI Overviews and AI Mode untouched, because those features run on the ordinary Googlebot crawl. The controls that do change what an AI answer can quote are the page-level robots rules: noindex removes the page, nosnippet withholds its text, and max-snippet caps how many characters may be shown or used. Set those in the meta tag or the X-Robots-Tag header, and treat Google-Extended as a separate decision about training.

By Gale Editorial · Updated 2026-09-15. Every figure cited to a dated source. How we write.

Does blocking Google-Extended keep a page out of AI Overviews?

No. Google-Extended is a robots.txt token, and Google's crawler documentation says it does not affect a site's inclusion in Search and is not a ranking signal 1. It governs whether crawled content may train future Gemini models and be used for grounding in Gemini Apps and on Vertex AI 1. AI Overviews and AI Mode are features of Search, and the robots.txt rules for Googlebot are the control for how a site is crawled for Search, AI features included 2.

Google-Extended has no user agent of its own; the crawling happens under the existing Google user agents, and the token works only as a control 1. So a practice that disallows Google-Extended across the whole site sees nothing change in its server logs and nothing change in Search. Googlebot keeps fetching the about page, and an AI Overview can keep quoting it.

That is the design. Google's AI features documentation says there are no additional requirements to appear in AI Overviews or AI Mode, and that the way to limit what they show from a page is the ordinary set of preview controls: nosnippet, data-nosnippet, max-snippet and noindex 2. The setting most owners reach for to manage AI is the one that changes nothing a searcher sees.

The three page-level rules that change what an AI answer can quote

Three robots meta rules decide it. noindex tells Google not to show the page in search results. nosnippet tells Google to show no text snippet or video preview, and the specification says it also keeps the content out of the direct input for AI Overviews and AI Mode. max-snippet caps the snippet at a number of characters, and the cap limits how much of the page may be used as direct input for the same two features 3.

RuleOn the results pageIn AI Overviews and AI ModeWhere it lives
noindexPage not shown in resultsNothing to quoteMeta tag or X-Robots-Tag header
nosnippetNo text snippet or video previewContent not used as direct inputMeta tag or X-Robots-Tag header
max-snippet:[number]Snippet capped at that many characters; 0 equals nosnippet, -1 lets Google chooseCap limits how much is used as direct inputMeta tag or X-Robots-Tag header

The special values deserve a second look, because a builder can set one without the owner knowing. max-snippet:0 is the same as nosnippet, and max-snippet:-1 lets Google choose the length 3. A cap of 50 is the quiet failure: the page stays indexed, it shows a fragment, and the fragment is all an AI feature may use. Outside those three sits data-nosnippet, a narrower attribute on a span, div or section that keeps one block out of snippets while the rest of the page stays quotable 3, the right tool for a fee that changes every quarter.

One more tag, noarchive, behaves differently at each engine. Google says it is no longer used by Search to control whether a cached link is shown 3. At Bing the same word still has teeth: a page carrying noarchive will not be linked in Chat and Copilot and will not train Microsoft's foundation models, and nocache limits Copilot to the URL, title and snippet 4. A noarchive tag left over from an old theme costs nothing on Google and costs the page its place in Copilot.

Where the rules hide: the meta tag and the X-Robots-Tag header

A robots rule lives in one of two places, and checking both on a small site takes under an hour. The first is a meta tag in the page's head named robots whose content is the rule; a user agent token in place of robots, such as googlebot, narrows it to one crawler 3. The second is the X-Robots-Tag HTTP header, and Google says any rule a meta tag can carry can also travel in that header 3.

Start with the source. Open the home page, the about page and the new-patient page, view the page source (Ctrl-U on Windows, Command-Option-U on a Mac), and search it for name="robots" and then for name="googlebot". Write down the content value on every hit. Then open the browser's developer tools, reload with the Network panel open, select the page's own request, and read the response headers for a line beginning X-Robots-Tag. A page can be clean in its source and still carry noindex in the header, which is how a rule set once on the whole server outlives the person who set it.

The same method is the raw-HTML test for everything else the page is supposed to say: if the hours, the address and the line about new clients are not in the source as text, an AI feature has nothing to quote. Site builders and plugins often set these rules from a checkbox worded as hiding the site from search engines, and a theme update can restore a default the owner switched off.

Which pages to open and which to keep closed

Open the pages a stranger would search for and keep closed the pages a client signs into. Google's optimization guide says that to be shown in generative AI features a page must be indexed, eligible to be shown in Search with a snippet, and the site must be included in Search generative AI features in Search Console 5. The about, services, fees and new-client pages should meet all three.

The portal, the intake form and any page holding a client's information keep noindex. For the pages you want quoted, remove nosnippet, take data-nosnippet off the blocks holding the plain facts (hours, address, whether new clients are accepted), and remove any max-snippet cap short enough to cut a sentence in half. A 40-word answer runs past 200 characters. A cap of 160 or less therefore shortens every passage worth quoting.

The optimization guide also says what you do not have to add. No llms.txt file, no AI text file, no Markdown and no special markup: Google says Search itself does not use them, so they neither help nor harm 5. For a solo practice the whole job is the public half of the practice website, indexed and unrestricted, with the plain facts as visible text. The website is one of the five sources an AI answer can draw from, and the one where a single tag can take it off the list. The five-page website a new practice starts with already has the right pages; the work is making sure no tag holds them back.

Deciding on Google-Extended as a separate question

Whether to block Google-Extended is a decision about training and grounding, and nothing about it changes what a searcher sees. Google's crawler list describes the token as a way to manage whether crawled content may train future Gemini models and be used for grounding in Gemini Apps and on Vertex AI, and states that it does not affect inclusion in Search or ranking 1.

The rule is two lines in robots.txt: a user-agent group named Google-Extended, followed by a disallow for the paths to keep out of training, the whole site or one directory 1. Google's robots.txt guide says the file manages crawler access and is not the mechanism for keeping a page out of Google; a disallowed URL can still be indexed from links elsewhere, and different crawlers may read the same rules differently 6. And the token binds one company. The crawler allowlist for every other engine is its own list of tokens, and none of them reads the Google-Extended line.

A practice that wants its public pages quoted and kept out of training can hold both positions at once: no nosnippet on the about page, and a Google-Extended disallow in robots.txt. The two settings do not touch.

How to see whether it worked

Search Console now shows the answer directly. In June 2026 Google introduced Search Generative AI performance reports, with impressions of a site's URLs inside AI Overviews and AI Mode broken down by page, country, device and date, rolled out to all sites worldwide by the end of August 2026 7. A page that carried nosnippet last month and has impressions there this month has crossed the line the tag was holding.

The report counts impressions and pages, and shows no queries or clicks. Read it alongside the monthly prompt panel: the same handful of questions a new client would type, asked each month, with a note of whether the practice's page was quoted. The report says Google could show the page; the panel says whether a question drew it. Give a removed tag time to be recrawled before judging either.

Gale publishes a provider's Q&A capsules on the gale.care profile page as visible text and announces that page through IndexNow the day it changes. The IndexNow setup does the same job for your own pages at Bing, where noarchive and nocache, and no Google-Extended line, decide what Copilot may quote.

Common questions

No. Google-Extended is a robots.txt token that manages whether crawled content may train future Gemini models and be used for grounding in Gemini Apps and Vertex AI. Google states it does not affect a site's inclusion in Search and is not a ranking signal. AI Overviews and AI Mode are Search features, run on the ordinary Googlebot crawl, and the page-level snippet rules are what limit them.

A value of -1 tells Google to choose the snippet length it believes is most effective, which is also what a page with no max-snippet rule gets. A value of 0 equals nosnippet. A practice page needs no cap at all on its public pages; the rule to watch for is a small positive number a builder set, since that cap also limits how much of the page an AI feature may use.

On Google it is inert: Google says noarchive is no longer used by Search to control cached links. On Bing it is not. Bing's documentation says a page with noarchive is not linked in Chat and Copilot and is not used for training, and nocache limits Copilot to the URL, title and snippet. If you want the page quoted by Copilot, take the tag off.

No. Google's optimization guide says you do not need new machine-readable files, AI text files, markup or Markdown to appear in Search or its generative AI features, because Google Search does not use them. Eligibility rests on three things: the page is indexed, it is eligible to be shown with a snippet, and the site is included in Search generative AI features in Search Console.

Yes, with data-nosnippet. It is an HTML attribute placed on a span, div or section, and Google excludes the marked text from snippets while the rest of the page stays eligible. Wrap the fee line alone, leave the hours, the address and the new-client line unmarked, and avoid the page-wide nosnippet, which withholds the whole page from AI Overviews and AI Mode as direct input.

Open the page in a browser, open the developer tools, switch to the Network panel and reload. Select the first request, the one for the page itself, and read its response headers for a line that begins X-Robots-Tag. Any rule a meta tag can carry can travel in that header, and a header set on the whole server applies to every page, including ones whose source looks clean.

Run your practice on Gale

The software is free. Gale earns one flat 3.5% all-in per paid transaction — only on transactions that actually pay. No subscription, no setup fee, no network cut.

Start or manage a practice →

References

  1. 1.Google (2026). List of Google's common crawlers. Google for Developers (Crawling infrastructure documentation). link ✓Google-Extended is a robots.txt control token with no user agent of its own that governs training of Gemini models and grounding in Gemini Apps and Vertex AI, and does not affect inclusion in Google Search or ranking; the robots.txt group syntax.
  2. 2.Google Search Central (2025). AI Features and Your Website. Google Search Central documentation (developers.google.com). link ✓AI Overviews and AI Mode are Search features with no additional requirements or special optimizations; Googlebot's robots.txt rules govern crawl access and the preview controls (nosnippet, data-nosnippet, max-snippet, noindex) limit what they show.
  3. 3.Google Search Central (2026). Robots Meta Tags Specifications. Google Search Central documentation (developers.google.com). link ✓The definitions of noindex, nosnippet, max-snippet (0 and -1), data-nosnippet on span, div and section, noarchive no longer used by Google Search, the robots versus googlebot name attribute, and the X-Robots-Tag header carrying any meta rule.
  4. 4.Microsoft Bing Webmaster Tools (2026). Robots meta tags and attributes that Bing supports - Bing Webmaster Tools. Bing Webmaster Tools Help Center. link ✓At Bing, noarchive keeps a page out of Chat and Copilot and out of foundation-model training, and nocache limits Copilot to URL, title and snippet.
  5. 5.Google Search Central (2026). Google's Guide to Optimizing for Generative AI Features on Google Search. Google Search Central documentation (developers.google.com). link ✓Eligibility for generative AI features requires an indexed page eligible for a snippet and a site included in Search generative AI features in Search Console; no llms.txt, AI text files, Markdown or special markup are needed.
  6. 6.Google Search Central (2025). Robots.txt Introduction and Guide. Google Search Central documentation (developers.google.com). link ✓robots.txt manages crawler access and is not the mechanism for keeping a page out of Google; a disallowed URL can still be indexed from links elsewhere, and crawlers may interpret rules differently.
  7. 7.Hillel Maoz and Moshe Samet, Google Search Console (2026). Introducing Search Generative AI performance reports in Search Console. Google Search Central Blog. link ✓Search Console's Search Generative AI performance reports, announced June 2026 and rolled out to all sites by August 31, 2026, show impressions inside AI Overviews and AI Mode by page, country, device and date, without queries or clicks.

https://www.gale.care/for-providers/aeo-google-extended-ai-overviews · 7 sources. Competitor details are cited to dated public sources and maintained as they change; figures are estimates, not commitments. Synthetic demonstration.

← Back to for providers

Findability, by specialty

How practices like yours get found in local search and AI answers — the honest playbook, per specialty.

SEO for private practices · SEO for AI search / answer engines (all verticals)