What is AI Citation Share, and why should every clinic care?

For most of the last 25 years, patients found their doctor through Google. They typed a question, scrolled the blue links, clicked one, judged the site, and either booked or kept looking. The work of being "found" meant ranking in those blue links.

In 2024–2026, a new layer crept in front of the blue links: AI chat. ChatGPT for the general consumer, Perplexity for the careful, Claude for the meticulous, Gemini for the Google-tied. Increasingly, when a patient asks "İstanbul'da iyi bir botoks doktoru kim?", the first thing they see is a 200-word AI-generated answer that cites some sources — and decisively excludes others.

If your clinic isn't in those citations, you are invisible to this layer. The blue links beneath might still rank you, but the patient has already formed an opinion.

The new metric

AI Citation Share is the percentage of probe queries — questions a real patient might actually ask — for which your clinic appears in at least one AI system's response.

It's coarse, but it's directional. A clinic with 5% citation share across ChatGPT + Claude + Perplexity + Gemini is barely visible. A clinic at 40% is the default answer.

Why traditional SEO doesn't cover it

Google's blue-link ranking favors authority signals (backlinks, domain age, Core Web Vitals) and intent-matched content. AI citation favors something subtly different:

• Structured data. Schema.org JSON-LD with MedicalWebPage + Physician + Article markup roughly doubles inclusion odds in AI answers vs unstructured HTML.
• Named-entity clarity. AI systems do entity resolution. If your clinic name appears next to a recognized specialty + a verified credential, the engine is more confident citing you.
• Citation hygiene. Pages with explicit citations (links to authoritative medical sources) get cited back more often. Engines treat citation-rich pages as "reliable nodes."
• Voice consistency. A physician whose published opinions are consistent across pages gets attributed more confidently than one whose tone is inconsistent.

How MedSelect measures it

Every night, MedSelect runs ~10 probe queries (per tenant) against ChatGPT, Claude, Perplexity, and Gemini. The response text is parsed for: (a) URLs whose host matches a MedSelect-network canonical domain, (b) physician names appearing in the response. Each match counts as a citation. The daily ratio (citations / probes) is the citation share for that tenant on that day.

We don't believe the absolute number means anything in isolation — but the trend, especially after a content sprint or a voice card refinement, tells you whether your editorial work is actually reaching the layer that now sits between your patient and your site.

What to do this week

If you run a clinic site and you want to start playing this game:

1. Add Schema.org JSON-LD to your top 10 traffic pages.
2. Make sure every claim has a citation to an authoritative source.
3. Audit your physician profile pages — are they explicit about credentials, specialty, and clinic affiliation?
4. Pick 10 queries you want to be cited on. Type them into ChatGPT today. Note which sources got cited.
5. Repeat the test monthly. The delta is your AI Citation Share trajectory.

Or, get in touch — that's what we do.