AI is changing medical SEO in two ways. Google now puts an AI-written summary above the results for many health questions, and fewer people click through when it appears. And patients ask ChatGPT, Claude and Gemini to recommend a provider, getting a short list of named practices with no page of blue links at all. The practices that show up are the ones with clear, trustworthy, well-structured information that these systems can find and cite.
The good news for clinics: the work that earns citations is the same work that earns rankings. Google says there are "no additional requirements to appear in AI Overviews or AI Mode" beyond being indexed and eligible for a snippet (Google Search Central). What changes is how you measure success.
What AI Overviews do to health searches
Google's AI Overviews reached 1.5 billion monthly users in 200 countries and territories by May 2025 (Google). They show most often on informational questions, which is the kind of search many health articles target: "what causes heel pain," "how long does strep last," "is a root canal painful."
Pew Research Center tracked the browsing of 900 US adults in March 2025, covering 68,879 Google searches:
| Finding | Number |
|---|---|
| Searches that produced an AI summary | 18% |
| Visits where users clicked a traditional result, with an AI summary | 8% |
| Visits where users clicked a traditional result, without an AI summary | 15% |
| Visits where users clicked a link inside the AI summary | 1% |
| Share of AI summary sources that were .gov sites | 6% (vs 2% in standard results) |
Source: Pew Research Center, July 2025. Pew's study covered all topics; it did not break out health queries.
For a clinic, this means symptom and "what is" articles will bring fewer clicks than they used to, even at the same rank. Searches with local or booking intent ("orthodontist near me," "urgent care open now") still lead with the map pack, which sends patients straight to a profile.
How Google handles health in AI Overviews
Google has said that "for topics like news and health, we already have strong guardrails in place," and in May 2024 it "launched additional triggering refinements to enhance our quality protections" for health (Google).
Those guardrails are not perfect. In January 2026, Google removed AI Overviews from queries such as "what is the normal range for liver blood tests" after a Guardian investigation found the summaries gave ranges that ignored factors like sex, age and ethnicity (TechCrunch). Expect Google to keep adjusting which health queries get AI answers.
How ChatGPT, Claude and Gemini recommend providers
A growing share of patients skip the search box and ask an assistant. KFF's 2024 poll of 2,428 US adults found 17% use AI chatbots at least once a month for health information and advice, rising to 25% among adults under 30 (KFF). OpenAI launched a dedicated health space in ChatGPT in January 2026 (OpenAI).
When someone asks "Who is the best pediatric dentist in Round Rock?" the assistant usually:
- Runs one or more web searches behind the scenes.
- Reads practice websites, directories (Healthgrades, Zocdoc, Yelp), review sites, local news and the Google Business Profile data it can reach.
- Returns a short list of named practices with a sentence on why each was picked, often citing the pages it used.
A practice gets named when the facts behind the question are easy to find and agree across sources: specialty, location, ages served, insurance accepted, hours, years in practice, and reviews. If your website says "pediatric dentistry" in an image and your directory listings disagree on your address, the assistant has little to work with.
What different engines read
| Engine | How it finds pages | What you control |
|---|---|---|
| Google AI Overviews and AI Mode | Google's own index | Normal Google SEO; page must be indexed and snippet-eligible |
| ChatGPT search | OpenAI's OAI-SearchBot plus search partners | Allow OAI-SearchBot in robots.txt (OpenAI) |
| Claude | Anthropic's Claude-SearchBot and Claude-User | Allow those agents in robots.txt (Anthropic) |
| Gemini app | Google's index and Gemini grounding | Google SEO; Google-Extended controls Gemini use and "does not impact a site's inclusion in Google Search" (Google) |
Security plugins and CDN rules sometimes block AI crawlers without the site owner knowing. The free AI crawler checker shows which bots your robots.txt allows.
What gets cited
There is no published formula, but the evidence points to a few consistent traits.
Pages that already rank, and pages that rank for related questions. Ahrefs found 38% of AI Overview citations came from pages in the top 10 for the same query, down from 76% in its earlier study; it suggests AI Overviews draw on results for related sub-questions too (Ahrefs, 2026). Ranking well across a topic matters more than ranking for one keyword.
Trustworthy health sources. In Pew's sample, .gov sites made up 6% of AI summary sources against 2% of standard results. That skew suggests AI summaries favour sources with clear authority. For a clinic, authority means credentialed authors, a medical reviewer and links to guidelines. See healthcare SEO best practices for the E-E-A-T checklist.
Quotable, specific facts. A sentence like "We see children from age 1 through 18 and accept Delta Dental, Cigna and Texas Medicaid" can be lifted into an answer. A paragraph of marketing copy can't.
Structured data. Schema.org markup (Dentist, MedicalClinic, Physician) states your name, address, specialty and hours in a machine-readable form. See healthcare schema markup.
Consistent third-party listings. Directory profiles, review sites and local news coverage give assistants confirming sources. See NAP consistency.
What to change on your site
| Change | Why it helps AI answers | Effort |
|---|---|---|
| Add a facts block to each location page: specialties, ages, insurance, languages, hours | Gives assistants quotable answers to provider questions | 1 hour a location |
| Add "written by" and "medically reviewed by" with credentials to clinical pages | Matches the trust signals Google asks raters to check | 1 day |
| Answer patient questions in short FAQ sections | Matches how people phrase questions to assistants | Ongoing |
| Mark up practice and providers with schema.org | Machine-readable facts | Half a day |
| Allow OAI-SearchBot, Claude-SearchBot and Googlebot in robots.txt | Engines can only cite what they can read | 15 minutes |
| Complete directory profiles with matching details | Confirming sources for recommendations | 1 to 2 days |
| Publish an llms.txt summary of the practice | Optional, low effort; adoption by AI engines is still limited | 30 minutes |
The free llms.txt generator drafts the file if you want to try it. For a broader view of AI search optimization, read what is generative engine optimization and how to get recommended by ChatGPT.
How to measure your visibility in AI answers
Organic rank and Search Console clicks can fall while your practice is cited more often in AI answers, or the reverse. Measure AI visibility on its own.
- List 10 to 20 patient questions that should lead to you: "best [specialty] in [city]," "[specialty] near [neighborhood] that takes [insurance]," "who treats [condition] in [city]."
- Ask them in ChatGPT, Claude, Gemini and Google AI Mode. Record whether your practice is named, which competitors are, and which pages are cited.
- Repeat monthly. Answers vary from run to run, so look at the share of questions where you're named over time.
- Fix the gaps. If an assistant names a competitor and cites a directory page, check your listing on that directory. If it cites a competitor's service page, compare it with yours.
The free Rank.ai AI grade does the first pass for you: it asks ChatGPT, Claude and Gemini 12 questions your patients ask, scores how often they name your practice out of 100, and lists the competitors named, the pages the AI read and three fixes. For daily tracking, Rank.ai's AI visibility tracking records who is named and which pages are cited every day.
Using AI to produce medical content
Many practices now use AI to draft content. Google's guidance allows it, with a warning that matters for health: "It is critical to manually factcheck and review all AI-generated content for accuracy and trustworthiness before publishing" (Google Search Central). Publishing many pages "without adding value for users" can breach Google's scaled content abuse policy.
For a clinic, the safe rule is simple: AI can draft, a qualified clinician reviews and signs every clinical page, and the page names that reviewer. Our healthcare content marketing guide sets out the full workflow, and AI SEO tools compares the tools.
FAQ
How is AI changing medical SEO?
AI Overviews now answer many health questions at the top of Google, which reduces clicks to websites: Pew found users clicked a result on 8% of visits with an AI summary against 15% without. Patients also ask chatbots to recommend providers. Practices now need to rank, be cited inside AI answers, and be named when assistants recommend local providers.
Do AI Overviews show for medical questions?
Yes, for many informational health questions, though Google applies stricter guardrails to health and adjusts which queries trigger them. In January 2026 Google removed AI Overviews from some liver-test queries after reporting found misleading answers. Searches with local intent, such as "dentist near me," still lead with the map pack.
How do I get my practice recommended by ChatGPT?
Make sure ChatGPT's search crawler (OAI-SearchBot) can reach your site, and publish clear facts about your specialties, location, insurance and hours on your pages. Keep directory listings consistent and keep earning reviews, because assistants cross-check several sources. Then test the questions patients ask and track whether you are named.
Does schema markup help with AI search?
Schema markup gives AI systems your name, address, specialty and hours in a structured form, which makes those facts easier to extract. Google says no special markup is required for AI Overviews. Treat schema as one part of making your information clear, alongside readable page text.
Should medical sites block AI crawlers?
Blocking training crawlers such as GPTBot or Google-Extended does not remove you from Google Search, and some practices choose to block them. Blocking search crawlers such as OAI-SearchBot or Claude-SearchBot can keep your practice out of those assistants' answers. Decide bot by bot, and check your robots.txt so a security rule hasn't blocked them by accident.
