How ChatGPT Decides Which Doctor to Recommend (And Why You're Probably Not on the List)
AI assistants are becoming the new front door for patient discovery. Here's exactly how ChatGPT, Claude, and Gemini decide which specialists to recommend — and what you can do about it.
Last month, a cardiologist in Madrid asked us a question that stopped us in our tracks: "I've been practicing for 22 years. I have a full patient load. But my daughter just told me she asked ChatGPT for the best cardiologist in Madrid, and my name wasn't in the answer. How is that possible?"
It's possible because AI doesn't know what you know. It knows what it can find.
The new front door of medicine
Something fundamental has changed in how patients find specialists. For two decades, the playbook was straightforward: rank well on Google, maintain good reviews, and patients would find you. That playbook still matters — but it's no longer the only one.
A growing number of patients now ask AI assistants for recommendations. They type "Who is the best dermatologist in Berlin?" into ChatGPT, Claude, or Gemini and expect a confident, specific answer. And they get one. The AI names names.
The question is whether your name is among them.
How AI actually decides
AI models don't have a secret ranking algorithm the way Google does. They work differently — and understanding the difference is the key to fixing your visibility.
When a patient asks ChatGPT for a specialist recommendation, the model does something remarkably similar to what a very thorough research assistant would do: it searches the web, gathers information from multiple sources, cross-references what it finds, and assembles an answer based on the evidence.
The critical word there is evidence. AI doesn't recommend doctors it "likes." It recommends doctors it can verify. And verification requires specific types of digital signals.
The six things AI looks for
Through our work auditing hundreds of medical professionals across Europe, we've identified six categories of signals that determine whether AI recommends you:
1. Does AI know you exist?
This sounds basic, but it's where many doctors fail. AI needs to confirm your existence through multiple independent sources. Your name, title, specialty, and city need to appear consistently across the web. If your name is spelled differently on your website, your LinkedIn, and your directory listing, AI may treat these as three different people — or worse, none of them confidently enough to recommend.
2. Do you show up in specialty searches?
There's a critical difference between AI knowing who you are and AI recommending you. A model might correctly identify you as a cardiologist in Madrid if someone asks about you by name. But when someone asks "Who is the best cardiologist in Madrid?" — a completely different query — you might be nowhere in the answer. This gap between recognition and recommendation is where most doctors lose patients.
3. Do authoritative sources vouch for you?
AI models are trained to trust certain sources more than others. A mention in a peer-reviewed journal carries more weight than a social media post. A feature in a national health publication signals more authority than a self-published blog post. A Wikipedia page — when legitimately earned — is one of the single strongest signals a doctor can have.
4. Are you in the directories AI checks?
AI models frequently reference structured medical directories because they're reliable, verified sources. Google Business Profile is disproportionately important for Gemini (which is built on Google's data). Doctoralia matters across Europe. Healthgrades matters in the US. If you're not in these directories, you're invisible to the models that rely on them.
5. Does your digital footprint reinforce your identity?
A professional website, an active LinkedIn profile, a presence on relevant social platforms — these don't individually make or break your score, but together they create a pattern that AI interprets as confidence. A doctor with a website, LinkedIn, and directory listings presents a consistent, verifiable identity. A doctor with nothing but a hospital bio page is a weaker signal.
6. Is your information accurate and current?
AI checks whether the information it finds is consistent and up to date. Old addresses, outdated specialties, or conflicting details across platforms reduce AI's confidence in recommending you.
Why each AI model gives different answers
One of the most surprising discoveries doctors make is that ChatGPT, Claude, and Gemini often give completely different recommendations for the same query. This isn't a flaw — it's because each model weights evidence differently.
ChatGPT gives outsized weight to Wikipedia and major English-language publications. If you have a Wikipedia page, ChatGPT is significantly more likely to recommend you. If you don't, you're fighting an uphill battle with this model.
Claude focuses on authoritative, institutional sources. Academic publications, professional association listings, and institutional team pages carry heavy weight. Claude is the hardest model to score well on without genuine authority signals.
Gemini is deeply tied to Google's own data. Your Google Business Profile, Google reviews, and Google Maps presence are disproportionately influential. A complete Google Business Profile with strong reviews can single-handedly make you visible on Gemini.
This means a doctor with a strong Google presence but no publications might score well on Gemini but poorly on Claude. Understanding this difference is the key to a targeted improvement strategy.
What most doctors get wrong
The biggest misconception is that AI visibility is just "SEO but for AI." It's not. Google SEO is about keywords, backlinks, and page speed. AI visibility is about identity verification and authority accumulation across multiple independent sources.
You can't keyword-stuff your way into a ChatGPT recommendation. You earn it by building a digital presence that gives AI enough evidence to confidently say: "This is a real, qualified, reputable specialist in this city, and I can recommend them."
What you can do about it
The good news is that the steps are concrete and achievable. Most doctors who are invisible to AI can reach meaningful visibility within 60-90 days by focusing on a few high-impact actions:
- Claim your Google Business Profile with complete, accurate details. This is the single highest-ROI action for most doctors.
- Create or optimize a personal professional website with your name, specialty, city, and credentials clearly stated.
- Ensure your LinkedIn headline includes your specialty and city — not just "Medical Doctor."
- Get listed on your region's primary medical directory (Doctoralia in Europe, Healthgrades in the US).
- Ask 5-10 satisfied patients to leave Google reviews mentioning your specialty.
These five steps alone address the most common gaps we see in doctor audits.
The window is now
AI-driven patient discovery is still early. Most doctors haven't heard of GEO yet, let alone optimized for it. That means the doctors who act now have a significant first-mover advantage — there are fewer competitors for AI recommendation slots than there are for Google search rankings.
But this window won't stay open. As awareness grows, the competition for AI visibility will intensify. The question isn't whether you need to think about this. The question is whether you want to be ahead of the curve or behind it.
Want to know where you stand? Run your AI visibility audit in under 3 minutes and see exactly what ChatGPT, Claude, and Gemini say about you.
Ready to check your AI visibility?
Find out if patients find you when they ask ChatGPT, Claude, or Gemini for the best specialist in your city.
Run Your Audit