According to Google, AI Overviews now appear on a majority of informational searches in the US, and health questions are among the categories where they show up most often. If a patient asks "what causes chronic migraines" or "best treatment for rosacea," there is a good chance they see an AI-generated summary before they see a single blue link — including yours.
That summary either names your clinic, or it doesn't. There is no middle ground, and no ad spend that buys your way in.
This guide covers what Google's AI Overview actually is, why it behaves differently from ChatGPT or Perplexity, and the concrete steps that improve your odds of being the source it pulls from — and cites.
What Google AI Overview Actually Is
AI Overview is the AI-generated answer block Google shows directly inside search results, above or alongside the traditional ten blue links. It launched in the US in May 2024 as the production rollout of what Google had been testing as SGE (Search Generative Experience) — if you see that older name in other articles, it's the same feature.
Functionally, it works differently from a chat assistant. AI Overview is generated per-query, in real time, from Google's live index — it is not primarily a conversation with a model trained on static data. That means the pages Google already trusts enough to rank tend to be the same pool it draws AI Overview citations from. Ranking well and appearing in AI Overview are related, but they are not the same outcome, and a page can do one without the other.
Why This Matters More for Healthcare Than Most Categories
Google treats medical queries as "Your Money or Your Life" (YMYL) content — a category where inaccurate information carries real-world risk. That classification means Google applies a higher bar for the sources it is willing to summarize and cite in an AI Overview for health questions than it does for, say, recipe or travel queries.
In practice, that higher bar comes back to the same signals search quality raters have used for years: expertise, experience, authoritativeness, and trust — E-E-A-T. Across the 595 clinic sites we benchmarked for our AI-visibility audit, E-E-A-T was the single weakest layer, averaging 34.82 out of 100 — meaningfully behind content quality (53.18) and technical readiness (45.48). That gap is exactly what determines whether Google trusts a clinic enough to summarize it.
There is also a business risk unique to AI Overview: because it appears above organic results, a well-answered query can become "zero-click" — the patient gets their answer and never visits any website at all, including the ones cited. If you're not the source AI Overview names, you're not just losing a click. You're losing the recommendation itself.
How Google Selects Sources for AI Overview
Google hasn't published an exact ranking formula for AI Overview citations, but the pattern across observed results points to a consistent set of factors:
- Existing organic strength. Pages that already rank reasonably well for the query, or closely related queries, are far more likely to be pulled into the summary.
- Structured, extractable answers. Content that states the answer directly and early — rather than burying it under three paragraphs of preamble — is easier for the system to lift cleanly.
- Schema markup. Structured data doesn't guarantee inclusion, but it gives Google a formal, unambiguous description of what a page and its author actually are, which lowers the risk of a citation being wrong.
- Author and site-level trust signals. For YMYL topics specifically, visible authorship, medical review, and credentials matter more than they do for lower-stakes categories.
- Crawl and index access. A page Google can't fully crawl or has deprioritized in its index isn't in the candidate pool at all, regardless of how good the content is.
None of these factors is exotic. What's notable is how rarely clinic sites get all five right at once.
Step-by-Step: Getting Your Clinic Ready
1. Check your current AI Overview visibility
Since May 2024, Google Search Console has included a dedicated AI Overviews filter under the Search type dropdown, showing impressions and clicks attributed specifically to AI Overview appearances. Most clinic accounts have never opened it.
Open Search Console, filter by Search type → Web, then apply the AI Overviews appearance filter. Note which queries already generate AI Overview impressions for your domain, and separately, run 10–15 of your highest-intent queries manually in Google to see whether a competitor is being cited instead of you. That comparison — Search Console data plus manual spot-checks — is your baseline.
2. Strengthen E-E-A-T before anything else
Given the YMYL bar, this is the highest-leverage fix for most clinics. At minimum:
- Every clinical article should show a named author and, where relevant, a named medical reviewer
- Author bios should state real credentials — degree, license, specialty, years in practice
- Claims that reference statistics or treatment efficacy should link to a source, not just assert the number
- The site should have visible, current privacy and data-handling pages, which Google's YMYL guidance treats as a trust signal for medical sites
If your physician pages are the weak point, our physician profile page guide covers the exact structure to fix it.
3. Implement complete schema markup
Prioritize, in this order:
MedicalOrganizationon the homepage, with name, address, phone, and hoursPhysicianon every doctor page, with credentials and specialtyMedicalProcedureorMedicalWebPageon service pagesFAQPageon any page with a genuine question-and-answer section
Validate with the Google Rich Results Test and the Schema Markup Validator (validator.schema.org) before treating any page as done. A full breakdown of this layer, including diagrams, is in technical GEO for medical websites.
4. Write for extraction, not just for ranking
AI Overview tends to lift self-contained answers. For your highest-priority pages:
- Answer the core question in the first sentence or two of a section — not after several paragraphs of context
- Use a genuine question as the heading when the page is answering one (this also feeds FAQPage schema)
- Keep each answer block understandable on its own, without requiring the reader to have read the paragraph above it
- Use comparison tables for anything involving options, pricing ranges, or timeframes — tabular data is easy for AI systems to lift accurately
5. Confirm Google can actually crawl the page
Check robots.txt for a Google-Extended user-agent rule — this specifically governs whether Google can use your content for AI features, separate from standard search indexing. A blanket Disallow here removes you from AI Overview consideration even if the rest of the page is perfect. While you're there, check for GPTBot, ClaudeBot, and PerplexityBot too, since the same legacy Disallow: / rules that block Google-Extended often block every other AI crawler at once.
6. Track it as its own metric
Re-check the Search Console AI Overviews filter monthly, not just organic rankings. A page can hold its ranking position while losing or gaining AI Overview presence independently — that shift won't show up anywhere else in your reporting.
Common Mistakes Clinics Make Here
The most frequent failure isn't a missing schema tag. It's treating AI Overview optimization as a one-time technical checklist instead of an ongoing trust-building exercise. Google re-evaluates which sources it's willing to summarize continuously; a clinic that added author bios once in 2024 and never touched them again is already behind a competitor who keeps credentials, reviews, and citations current.
The second most common mistake is optimizing content for AI Overview while leaving crawl access misconfigured — fixing the page but never checking whether Google-Extended can actually reach it.
What This Doesn't Replace
AI Overview is one surface among several. Optimizing for it does not substitute for answer engine optimization across ChatGPT and Perplexity, which weigh sources differently and don't share Google's index. Nor does it replace an ongoing content strategy — see clinic content strategy for AI search for the formats that consistently get cited across engines, not just Google's.
Treat AI Overview as one measurable layer of a broader AI-visibility practice, not the whole practice.