Ask an Assistant Who the Best Practice in Your City Is. Then Read the Answer Properly.

Here is a test you can run before you finish this article, without a vendor and without spending anything.

Open ChatGPT. Ask it who the best practice in your city is for the thing you most want to be known for. Use the words a patient would use, not the words you would use.

"Best plastic surgeon in Bellevue for a facelift"

"Who should I see for veneers in Seattle?"

"Best orthodontist near Kirkland for a teenager"

Then ask it two more times with slightly different wording, and once in Perplexity.

Most practice owners have never done this. That is worth sitting with for a moment, because it is a question a patient can ask about your business in five seconds and you have never seen the answer.

The four outcomes, and what each one means

1. You are named.

Good, and less final than it feels. Ask three more variations before concluding anything. Being named in one phrasing and absent in three others is the common pattern, and it means your presence is unstable rather than established. Note also which practices are named alongside you, because that list is your actual competitive set as a system understands it, and it is frequently not the list you would have written.

2. A competitor is named and you are not.

This is the finding people expect and dread. It is also the most actionable, because you can now go and look at what that practice publishes that you do not. In our experience the differences are rarely mysterious. The named practice usually has clearer service pages, more consistent information across directories, more review depth, or simply answers the cost question in public while you do not.

3. You get a directory, an aggregator or a hospital system instead of any practice.

Very common, and better news than it looks. It usually means no practice in your market has established itself clearly enough to be named, so the system falls back to a list. An unclaimed position is considerably cheaper to take than one a competitor already holds. If you are going to act on any of these outcomes, this is the one that rewards moving quickly.

4. The assistant declines to recommend anyone.

Also common in health contexts, and not a failure of your marketing. These systems are cautious about medical recommendations. When this happens, change the question. Ask what to look for when choosing a provider, or ask about the procedure rather than the person, and see whether your practice appears as a source in the explanation. Being the source a system draws on is how you get named in the follow up question.

Now the harder question: is what it says about you accurate?

Ask directly. "Is [your practice name] a good [specialty] practice?" or "Tell me about [your practice]."

This is the query almost nobody measures and the one with the most exposure. When a patient hears your name from a friend, this is the credibility check they run, and the answer is assembled from whatever is findable and corroborated about you. It will not appear in your analytics. You will never know it happened.

What we tend to find when practices run it:

• Stale information. A provider who left two years ago. A former address. An old practice name from before a rebrand.
• Merged identities. Your practice conflated with a similarly named one in a nearby city.
• Thin answers. A generic description that could apply to any practice in your specialty, because there is nothing distinctive and verifiable to say.
• Occasionally, something wrong. Not malicious, just assembled from a directory profile nobody has updated since 2019.

None of this is fixed by appealing to the system. It is fixed by correcting the sources it drew from.

Why the answers keep changing

If you run the same question five times you may get five different answers. That is normal, and understanding why it happens is most of what separates a useful test from an anxious one.

Answers vary with the exact phrasing, with location context, with account history, with the specific product and version, and with the date. There is no stable ranking underneath to be number one in.

The variation is itself a finding. Wide variation across phrasings usually means nobody in your market has established a stable presence. Narrow variation, with the same two or three practices appearing consistently, means somebody has, and you are either one of them or you are working against an incumbent.

What this test cannot tell you

Being honest about the limits, because the category is not.

It does not tell you how many patients are actually asking. We do not know that number for healthcare and neither does anyone quoting you one. Every percentage in circulation traces back to a vendor selling the service, and we would rather say so than repeat it.

It does not tell you whether being named produces bookings. That requires attribution most practices do not have.

It is not a benchmark. One person running six questions on one afternoon is a spot check, not a measurement. A measurement fixes the question set, names the systems, records the date and location context, and gets re-run identically later so the comparison means something.

What the spot check is genuinely good for is deciding whether to care. If you are named consistently, you can put this down. If a competitor is named consistently, or the answer is a directory, you now know something about your market that you did not know ten minutes ago, and it cost you nothing.

If you want to be more rigorous about it

Write down twenty questions a real patient would ask across your main procedures. Include the cost questions, because patients ask those first and most practices never answer them. Run all twenty in ChatGPT, Perplexity, Google AI Overviews and Gemini. Record for each: were you named, who else was, and what the date and location context were.

That is a baseline. Repeat it in ninety days with the identical questions and you have a measurement rather than an impression.

You can do this yourself. We would rather you ran the short version before talking to anyone, including us, because a practice that has seen its own answer asks much better questions of a vendor than one that has not.