If you took your phone out right now and Googled “best cardiologist near me,” you’d likely get an AI Overview with a snippet showing a handful of recommended doctors, their credentials, affiliations, and overall patient sentiment.
But how do Google, ChatGPT, Claude, and other emerging Large Language Models (LLMs) determine which doctors appear in these results? Why does each AI tool sometimes provide different answers?
Practices need to fully understand where AI gets doctor information and the distinct way it will be presented. This is the only way to adapt to the way patients now evaluate options for care.
Understanding the Universal “Core” Sources
AI has become the path of least resistance when prospective patients search for care. This change in search behavior means there are now even more opportunities for your practice and provider information to appear — which makes accuracy and consistency more important than ever. Google AI Overviews, Gemini, ChatGPT, Copilot, oh my!
To track this change more closely, we conducted an experiment asking five major LLMs — Gemini, ChatGPT, Perplexity, Claude, and Microsoft Copilot — the same four physician-related questions twice:
good surgeon?"
We found that while each AI has a unique “personality,” they all rely on a shared foundation of data.
So where does the data come from? Well, largely from the “big three” directories: Healthgrades, Vitals, and WebMD. All three are healthcare-specific directories full of provider bios, hours, locations, and verified patient feedback. AI models also frequently cross-reference data using practice websites to verify clinical credentials and details.
Getting to Know LLM Personalities
Beyond the core directories, each LLM has a distinct strategy for gathering deeper patient sentiment and local data.
Take a closer look at the sources each model uses most:
| Source | Gemini | ChatGPT | Claude | Perplexity | Copilot |
|---|---|---|---|---|---|
| Google Reviews | ✔ | ||||
| rater8 | ✔ | ✔ | ✔ | ✔ | ✔ |
| Healthgrades | ✔ | ✔ | ✔ | ✔ | ✔ |
| Vitals | ✔ | ✔ | ✔ | ✔ | ✔ |
| WebMD | ✔ | ✔ | ✔ | ✔ | ✔ |
| Practice websites | ✔ | ✔ | ✔ | ✔ | ✔ |
| U.S. news sources | ✔ | ✔ | ✔ | ✔ | |
| ZocDoc | ✔ | ✔ | |||
| RateMDs | ✔ | ✔ | |||
| ✔ | ✔ | ||||
| YouTube | ✔ | ✔ | ✔ | ||
| Yelp | ✔ | ✔ | ✔ | ||
| Doximity | ✔ | ✔ |
← Swipe to see all sources →
| LLM | Google Reviews | rater8 | Healthgrades | Vitals | WebMD | Practice websites | U.S. news sources | ZocDoc | RateMDs | YouTube | Yelp | Doximity | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Gemini | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | |||
| ChatGPT | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | |||||
| Claude | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ||||
| Perplexity | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ||
| Copilot | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ |
What Stands Out:
Beyond the core directories, each LLM has a distinct strategy for gathering deeper patient sentiment and local data.
Take a closer look at the sources each model uses most:
1. ChatGPT
ChatGPT pulls from the most diverse range of secondary sites. In addition to the “big three” directories, it frequently cites: Doximity, Medical News Today, Healthline, Everyday Health, and Mapquest.
2. Claude
Claude showed a unique reliance on transactional appointment and rating platforms. It is most likely to prioritize data from: Zocdoc, RateMDs, Sharecare, and Wellness.com.
3. Perplexity
Perplexity is the model most likely to include non-traditional community sources. It pulls reputation data from Facebook and YouTube, in addition to local and speciality-specific publications.
4. Microsoft Copilot
Copilot heavily favors specific authority sites like rater8 Verified Reviews listings, Vitals, and authorized U.S. news sources such as Reuters, USA TODAY Network, and Financial Times. However, it has a notable quirk to keep in mind: preferring high ratings over volume. In our research, we’ve found that it recommended a doctor as “top-rated” with a 5-star rating based on only 5 reviews. Thus, Copilot factors in star ratings much more heavily than the number of reviews.
5. Gemini
Because it’s built by Google, Gemini is notably the only LLM that natively integrates results from Google AI Overviews, Google Reviews, and Google Maps.
The Rise of Voice Search
In the midst of emerging AI models, the way patients search is also shifting from typing to talking.
In a recent survey we conducted of over 1,000 U.S. patients, we learned that 25% of people have used voice search when looking for information about a doctor. But which LLM answers that voice query? It depends on the device, and in some cases, the user’s settings.
Android and Google Home devices use Gemini by default, Google’s generative AI model. Apple recently announced they’re partnering with Google’s Gemini, while also offering ChatGPT as an optional integration through Apple Intelligence for users who enable it.
Meanwhile, Amazon’s latest assistant, Alexa+, reportedly uses Claude alongside other generative AI models.
The takeaway here is that, for now, there’s no single gatekeeper for voice search. Much like searching on a desktop or phone, a patient asking Siri, “Who is the best dermatologist near me?” might get an answer completely different from someone who uses an Amazon assistant or Android.
How to Manage Your Online Reputation in the AI Ecosystem
Maintaining your online image is already a challenge, and AI has made it even more complex. When a patient is looking for answers, AI isn’t just “guessing”; it’s a sophisticated aggregator of your existing digital footprint.
The secret strategy to staying competitive and maintaining consistency across LLM- and AI-generated search results is an online presence that’s credible, current, and backed with authentic patient sentiment.
To start, aim to keep business hours, addresses, and provider information up to date on the core directories like Healthgrades, Vitals, and WebMD. And don’t overlook your social presence. LLMs like Perplexity pull from Facebook and YouTube, which means your social channels are now AI assets, not just engagement platforms.
But the reality is, the AI ecosystem is moving quickly, and keeping up with every platform, algorithm, and emerging search behavior trend isn’t realistic for most practices. You’re busy delivering care, not monitoring which LLM is pulling from which directory this week.
That’s why we built raVE, the rater8 Visibility Engine.
raVE offers AI-readiness and generative engine optimization (GEO) by building listings with structured data, compiling patient feedback into search-friendly formats, and generating consistent, keyword-rich reviews that help AI understand and trust your practice. No matter where patients begin their search journey, they’ll find the same trustworthy story about your practice.
Improve your visibility in AI-driven search.
See how raVE keeps your reputation, listings, and patient feedback aligned so patients see a consistent story everywhere they search.
