Why doesn't ChatGPT name my practice?

I’m a marketer, not a lawyer, and nothing here is legal advice. Anything that touches patient information, including how you reply to reviews, is a question for your compliance officer. This page covers what AI assistants read about a practice and what you can check without touching patient data.

A patient who knows your name and one who doesn’t

A returning patient asks “What are the hours for [practice]?” A new one asks “My knee has hurt for three weeks. What should I do, and who should I see near me?” The second question is where new patients come from, and it’s the one where AI answers are least likely to name you.

In my AI answers field report, which pools graded answers about 9 businesses in different industries, not medical practices specifically, the business was named 90.2% of the time when the question named it. On problem questions that named nobody, it was named 2.6% of the time (16 of 623 answers), and 6 of 9 businesses were never named once.

Where a named practice comes from

A symptom question can get a health explanation and “see a specialist” with no name at all, so note in your log whether an answer names any practice. When an answer does name practices, it has to get them from places that list doctors:

• Doctor directories. Healthgrades and Vitals, which list doctors whether or not they’ve claimed a profile, and booking sites like Zocdoc for practices that list there.
• Your Google Business Profile and the map listing behind it.
• Reviews on the platforms the answers quote.
• Insurer and hospital directories that list which doctors take which plans.

Your website explains your services. The assistants use it to describe what you do. They look elsewhere to decide who to name.

The root record: your NPI

Doctor directories compile profiles from public records, and the NPI registry is the record a HIPAA-covered provider has to keep current. Under HIPAA’s administrative rules, a covered provider must report changes to its NPI data “within 30 days of the change” (45 CFR 162.410, checked October 27, 2026). The record includes the practice location address and phone number (NPPES data fields, checked October 27, 2026).

Whenever a provider’s address or phone changes, check that provider’s NPI record before anything else.

What to check first

I don’t have a measured medical effect. The figures below are pooled across 9 businesses in different industries, not medical practices specifically. From them, I’d check these in order:

1. Facts that agree everywhere. In the shorter snapshot checks, 4 of 4 businesses had facts that disagreed across their own site and listings, at least 6 each. For a practice, check phone numbers, hours, accepted insurance and provider credentials first.
2. Accuracy when patients ask by name. Between 15% and 35% of the answers that named a business got at least one fact wrong (per-business range). A wrong phone number or the wrong insurance list reaches a patient who was about to book.
3. Reviews on the platforms the answers quote. 4 of 4 snapshot businesses had fewer Google reviews than the median competitor the answers named instead. That’s a pattern, not a proven cause. Before you ask patients for reviews or reply to one, settle with your compliance officer what you can say without confirming someone is a patient.
4. Claimed directory profiles. The medical directory guide covers claiming Healthgrades, Vitals and Zocdoc.

What doesn’t explain it

Blocked crawlers, almost never. In the field report data, pooled across 9 businesses in different industries, not medical practices specifically, all 530 fetches I made with AI and search crawler user agents across 9 sites came back HTTP 200.

One answer isn’t a pattern. The same question asked twice gets different practices. Ask three times and count.

Check what AI says about your practice

The free What AI says about you template has ten questions, a log and a fact ledger. For a practice:

• Fill questions 1 to 6 with the practice name, and repeat question 3 for each location.
• Write question 9 the way a new patient describes the symptom, not the way a clinician names the condition.
• Add a row to the ledger for each provider: name, credentials, locations and NPI address.

Ask in a private window, three times on each assistant your patients use, and keep any patient detail out of the questions. If the practice runs Google Ads, the conversion tracking guide covers measuring booked patients without sending health information.

When to bring me in

Bring me in when AI answers get your hours, address or services wrong, when they name competitors for the work you want, or when you’ve fixed your listings and want to know whether the answers changed.

I'm Conner Crowe, and I write every audit myself. Send your website and your market. There’s nothing to grant, and within 48 hours you'll have a written read on what each AI assistant says about you and where the wrong facts come from. No call needed, and nothing in your account or store changes.

Get a free AI answer snapshot →

Or run the check yourself with the free What AI says about you template.