Google AI Overviews and AI Mode for medical practices: what Google says and what to fix on your site

When a patient types a question into Google, the answer at the top of the page may now be written by AI: an AI Overview on the regular results page, or a longer conversation in AI Mode. Both show links to the websites they draw from. For an independent practice, that raises practical questions. Can your site be one of those links? Do you need new markup, a special file or an “AEO” package to get there? Can you keep your content out? And what happens when the answer pulls an old fact, like a plan you no longer accept, from your own pages?

This guide answers those questions from Google’s own documentation, checked on the date below. It covers how the features choose pages, what Google says you can skip, the snippet controls and the newer Search Console setting, how to measure what you get, and the practice facts worth fixing first. Rules for writing marketing content with AI tools are covered in our guide to AI-generated content in healthcare marketing. Structured data is covered in schema markup for practice websites, and clinical review of health pages in running a medical review program. For assistants outside Google, such as ChatGPT, Claude, Perplexity, Copilot and Siri, see our AI search visibility guide.

The short version

• There is no separate entry ticket. Google says a page needs to be indexed and eligible to show with a snippet in regular Search to be eligible as a link in AI Overviews or AI Mode, and that there are no additional technical requirements.
• You do not need special files or markup. Google says it does not use llms.txt or other AI text files for Search, and that there is no special schema.org markup for these features.
• Unique, accurate content is what Google asks for. Its guide says useful, non-commodity content will likely matter more over time than any other suggestion in it.
• You have controls. nosnippet, data-nosnippet, max-snippet and noindex limit what can be shown, and since August 31, 2026 Search Console has a site-level setting to exclude a site from Search generative AI features.
• Measure it in Search Console. AI Overviews and AI Mode impressions are counted in the Performance report and broken out in the Generative AI performance report.
• Fix your facts first. Hours, locations, insurance, services and prices on your site and Business Profile are the details an AI answer can repeat, and wrong ones are still your claims.

How AI Overviews and AI Mode choose pages

Google describes both features as built on its regular Search systems. Its guide to optimizing for generative AI features names two techniques. The first is retrieval-augmented generation, also called grounding: the system uses Google’s core ranking systems to retrieve relevant pages from the Search index, reviews the information on them, generates a response and shows links to pages that support it. The second is query fan-out: the model issues a set of related searches on subtopics of the question to fetch more results.

Two points follow for a practice. First, a page that does not rank or is not indexed in regular Search has no separate path into these answers. Second, a patient’s question can be split into smaller questions, so a page that answers one narrow part well, such as what a first visit involves or which plans you accept, can be relevant even if it is not about the whole question.

Google also says that meeting all requirements does not mean a page will be crawled, indexed or shown, and that AI Overviews appear only when its systems decide they add something beyond classic results. Anyone who promises that your practice will appear in an AI Overview is promising something Google itself does not.

What Google says you do not need

Google’s May 2026 guide has a section titled “Mythbusting generative AI search: what you don’t need to do.” It is worth reading before you buy any add-on labeled AEO (answer engine optimization) or GEO (generative engine optimization). Google says that from its point of view, optimizing for generative AI search is optimizing for the search experience, “and thus still SEO.” The guide lists these as things you can ignore for Google Search:

• llms.txt and other AI files or markup. Google says Search does not use them. You may keep one for other services; it will neither help nor harm your Google visibility.
• “Chunking” pages into small pieces. Google says its systems can understand more than one topic on a page, and that there is no ideal page length.
• Rewriting pages just for AI. Google says its systems understand synonyms and general meaning, so you do not need to cover each long-tail variation.
• Seeking inauthentic mentions. Google says its core ranking systems focus on high-quality content while other systems block spam, and the AI features depend on both.
• Overfocusing on structured data. It is not required for these features, and there is no special schema.org type to add, though it still helps with rich result eligibility.

The same guide warns against making a separate page for each way a question might be asked, including fan-out queries, mainly to influence rankings or AI answers. Google says that violates its scaled content abuse spam policy. It also says to be wary of third-party tools that promise ranking success or claim to use internal Google metrics, because no third-party tool has access to Google’s internal ranking or AI systems.

What Google says does help

The first item in Google’s guide is to create valuable, non-commodity content. It describes commodity content as material based on general knowledge that anyone could write, and non-commodity content as content with expert or first-hand insight that goes beyond the ordinary. It says to write from what you know, not to recycle what is already online or what a generative AI model could easily produce.

For a practice, the raw material for that is already in the building: how your office handles a first visit, what your clinicians explain to patients before a procedure, what your intake asks and why, and what a patient should bring. A page that explains those things in your own words, reviewed by a clinician, is the kind of page Google describes. A page that repeats a generic definition of a condition is not.

Google’s guide also lists technical basics that still apply: let Googlebot crawl the site in robots.txt and through any CDN or host, link important pages internally, keep content in text rather than only in images, give visitors a good experience on all devices, make structured data match the visible text, and keep Business Profile information up to date. It adds that Google Business Profiles can help a local business appear in AI responses and other results.

Health topics carry a higher bar in Google’s quality framework. Google’s Search Quality Rater Guidelines treat health as a “Your Money or Your Life” topic, and section 3.4.1 says that where a page on a clear YMYL topic gives information or advice, “a high level of expertise may be required for the page to be trustworthy.” Rater guidelines do not directly rank pages, but they describe what Google’s systems aim to reward. Our medical review guide covers how to put that expertise on the page.

Your practice facts are the first job

An AI answer about a local practice can draw on your website, your Business Profile and other pages about you. When those sources disagree, a patient can read the wrong one at the top of the page. The fix is not an AI tactic; it is a fact audit.

• Hours. Google Business Profile guidelines ask for your regular customer-facing hours and allow special hours for holidays and closures. Google also says businesses that operate only by appointment should not provide hours. Match the hours on your site to the profile, and update both when they change.
• Locations and names. The guidelines ask that the business name reflect the real-world name used on your storefront, website and stationery. A closed location left on a page of your site is a fact a search answer can still find.
• Insurance. Keep one current list of plans on your site, with the date it was checked, and remove plans you have left. Our insurance accepted page guide covers the wording.
• Services and providers. Remove services you no longer offer and providers who have left, or mark the pages clearly.
• Prices. If you publish cash prices, date them and keep old price pages from contradicting the current one.

Accuracy matters for more than visibility. Under section 5 of the FTC Act (15 U.S.C. 45), deceptive acts or practices in commerce are unlawful, and claims on your own website about insurance, prices or results are your advertising. An AI feature quoting an outdated page does not change who published it.

Controlling what Google shows from your pages

Google says AI is built into Search, so robots.txt rules for Googlebot are the control for crawling, and blocking Googlebot affects the page in all of Search, not only in AI features. For finer control, Google points to four existing preview controls:

• nosnippet. Google’s robots meta tag specification says it stops a text snippet or video preview for the page in all forms of search results, including AI Overviews and AI Mode, and prevents the content from being used as a direct input for those features.
• max-snippet:[number]. Limits the snippet to that number of characters across result types, including AI Overviews and AI Mode, and limits how much of the content may be used as a direct input for them. Google notes the limit does not apply where a publisher has separately granted permission, such as through in-page structured data or a license agreement.
• data-nosnippet. An HTML attribute that keeps a specific section of a page out of snippets while the rest of the page can still be shown.
• noindex. Keeps the page out of Google Search entirely.

Google’s specification adds that where rules conflict, the more restrictive one applies. Its troubleshooting advice is to check with the URL Inspection tool that Googlebot sees the control, and to allow time for a recrawl, which Google says can take from days to months.

Think before using these on main pages, since a page without a snippet also loses its regular search description. They can make sense for a page whose details change daily, or a section such as a staff-only note that should not be quoted out of context. Pages that should never be public, such as anything with patient information, belong behind a login, not behind a snippet tag.

The Search Console setting and Google-Extended

Search Console now has a site-level Search generative AI control, under Settings. Google’s help page says it was rolled out to all websites worldwide as of August 31, 2026, and that it covers AI Overviews, AI Mode and generative AI features in Google Discover. Including the site is the default. If you exclude it, Google says links to your site and its content will not appear in those features and its content will not be used as an input to generate AI responses there, while content from other sites will still be available and may look similar to yours. Google says the control is not a ranking signal for other parts of Search, and that a change generally takes a few days to take effect.

Google-Extended is a different control. Google’s crawler documentation describes it as a robots.txt token that manages whether content Google crawls may be used to train future Gemini models and for grounding in Gemini Apps and Vertex AI, and says it does not affect inclusion in Google Search and is not a ranking signal. On October 30, 2026 Google also announced plans to support a shared AI training preference in robots.txt, based on work at the IETF, with support aimed for 2027.

For a practice that wants new patients to find it, excluding the site from AI features is a trade-off: fewer places your pages can appear, and other sites’ content in the answer instead. Decide it on purpose, record who decided and why, and recheck it if Google changes how the control works.

Measuring AI Overviews and AI Mode

Google says appearances in AI features count toward overall traffic in the Search Console Performance report, under the Web search type. Since August 31, 2026 there is also a Generative AI performance report that shows impressions from AI Overviews and AI Mode by page, country, device and date. Google’s help page says the report may not appear for a site that has not received enough impressions in these features, or that has excluded itself.

Two cautions. The report counts impressions, so a rise does not by itself mean more calls or bookings; track those in your own systems, such as booked appointments by source. And Search Console data for a small practice is thin, so read it over months, not days. If you use analytics on your site, keep patient health information out of it; our tracking pixels guide covers why.

Questions to ask an agency selling AI search work

• Which of the tasks you propose does Google’s own generative AI guide list as unnecessary, such as llms.txt files or chunking pages?
• Will you create pages for variations of the same question? How is that different from what Google calls scaled content abuse?
• Where do your visibility numbers come from? Google says no third-party tool has access to its internal ranking or AI systems.
• Will you seek mentions or links on other sites? Are any of them paid or arranged?
• Who writes and who clinically reviews health content you produce?
• Will you check our hours, locations, insurance list and providers on the site and the Business Profile before writing anything new?
• Will you change our Search generative AI control or robots.txt, and who signs off?

Use and avoid

• Use: “We are building pages that answer the questions our patients ask, reviewed by our clinicians.” Avoid: “We will get you into AI Overviews.”
• Use: “Accepted plans, checked [date].” Avoid: “We take all major insurance.”
• Use: “Hours updated [date]; holiday hours below.” Avoid: hours on the site that differ from the Business Profile.
• Use: one clear page per real service. Avoid: dozens of near-copies for each phrasing of a question.
• Use: Search Console impressions alongside booked appointments by source. Avoid: “AI visibility scores” from tools that claim to see inside Google.

An illustration

This example is invented. [Practice name], a family dental office in [Town], dropped one dental plan in the spring and updated its insurance page. A front desk team member mentions that two callers this month asked about the old plan and said they had seen it “on Google.” The office checks its site and finds an older new-patient page that still lists the plan, plus a blog post from two years ago that mentions it.

The office does not buy an AI add-on. It updates the new-patient page to point to the single insurance page, adds a dated note to the old blog post, checks that the Business Profile matches, and asks Google to recrawl the changed pages. Over the next months it watches the Generative AI performance report for those pages and keeps counting calls about the dropped plan. Whether the answers change depends on Google, but the office’s own sources now agree.

Checklist

• Confirm key pages are indexed and show a snippet in regular Search.
• Check robots.txt and any CDN or host settings do not block Googlebot.
• Audit hours, locations, providers, services, insurance and prices across the site and the Business Profile.
• Remove or date pages with outdated facts; keep one source page for insurance and one for prices.
• Write service pages from first-hand practice knowledge, reviewed by a clinician.
• Make sure structured data matches the visible text.
• Decide on purpose whether any page needs nosnippet, max-snippet or data-nosnippet.
• Record the Search generative AI control setting in Search Console and who chose it.
• Review the Generative AI performance report and booked appointments by source each quarter.
• Turn down tactics Google lists as unnecessary or spam.

How we help

We build practice websites, local SEO, Google Ads and lead generation and an AI front desk for independent practices. For search, including AI Overviews and AI Mode, we start with the facts on your site and Business Profile and with pages written from your clinicians’ knowledge, and we do not sell the AI files or tactics Google says it does not use. Our tools are built with HIPAA in mind. We confirm in writing what a BAA covers before any patient information flows through a tool, and if it is not covered, we do not route patient information through it. We are not a healthcare provider and we do not give legal or medical advice. For the wider rules, see our healthcare marketing rules guide.

Sources

• Google Search Central: AI features and your website
• Google Search Central: Optimizing your website for generative AI features on Google Search
• Google Search Central Blog: Top ways to ensure your content performs well in Google’s AI experiences on Search (May 2025)
• Google Search Central: Robots meta tag, data-nosnippet and X-Robots-Tag specifications
• Google Search Central: Google crawler list (Google-Extended)
• Google Search Central Blog: Supporting a shared AI training control for websites (October 30, 2026)
• Search Console Help: Search generative AI control
• Search Console Help: Generative AI performance report (Search)
• Google Search Central: Creating helpful, reliable, people-first content
• Google Search Quality Rater Guidelines
• Google Business Profile Help: Guidelines for representing your business on Google
• 15 U.S.C. 45: Unfair methods of competition unlawful (FTC Act section 5)

More in Local SEO and reviews for practices