Gemini names a dentist for nervous patients, ChatGPT rarely does

Short answer: ask Gemini for the best dentists in your city, and it will very likely tell you which one to see if you are scared. In 137 of 178 answers, collected in 89 cities across eight countries, it tied fear of the dentist to a practice it named: "Best for: Nervous patients". ChatGPT did that in 3 of its 178 answers. When it raised fear at all, 12 times, it was mostly general advice: look for a practice that offers sedation.

For a nervous patient, these are two different experiences. One assistant hands over a name with a promise of gentleness attached. The other says what to look for and leaves the choice to you. And Gemini's promise was written without reading a single review that day: it answered without web search.

The customer asks for the best dentist. Gemini also answers their fear, and gives it a name.

What was measured

Eight Waseit studies asked the same two buyer questions about dentists - the best dentists in a city, with a list of recommendations, then the most recommended ones and why - in 89 cities: 13 each in France (30 July 2026), the United Kingdom (31 July), Brazil (6 August) and the United States (13 September), then 11 in Canada, 10 in Australia, 6 in New Zealand and 10 in Singapore (16 October). The questions were asked in the local language: French in France, Portuguese in Brazil, English elsewhere. ChatGPT (gpt-4o-mini) and Gemini answered through their programming interfaces, without web search: 178 answers each. Perplexity, which searches the web by design, was asked in France, the UK and Brazil: 78 answers. We pulled every sentence that speaks of fear of dental care - anxious or nervous patients, dental phobia, fear of the dentist, sedation - and read all 159 passages. Two spoke of sedation for surgery, not fear, and were set aside. Each answer was then classed: does it tie fear to a practice it names, or only give general advice?

Which assistant tells a nervous patient where to go?

Gemini does, nearly every time, and by name. In the six English-speaking countries we measured, it tied fear of the dentist to a named practice in 120 of 126 answers: "Best for: Nervous patients", "excellent at putting nervous patients at ease", "They offer sedation for anxious patients". In French it did so in 11 of 26 answers, in Portuguese in 6 of 26, often for children. In a little over half of these answers, the same sentence credits patients' reviews or the practice's reputation. Yet Gemini answered without web search: no review was read that day. ChatGPT almost never does. It spoke of fear in 12 of its 178 answers, and 9 of those were general advice, such as looking for "options for sedation dentistry if needed". Three times it named a practice for it. Perplexity, which searched the web, raised fear in 2 of 78 answers, both times about a named practice. So the same question gets a name and a promise of gentleness from one assistant, and a checklist from the other. And the names themselves are not always right: when AI names a dentist, one in five is wrong.

What it means if you are the practice

If you look after nervous patients well, the assistant that names you for it is not reading your page that day: Gemini, without search, repeats what it absorbed earlier. So say it in text, on your site, in your patients' language: what you do for them (longer first appointments, explaining each step, sedation if you offer it and who provides it), and how to ask for it when booking. Do not claim what you do not offer: an assistant that repeats a claim turns it into a promise. When Gemini does search, as in our audit, a clear page is the kind of thing it can read. These are reasonable steps, not levers with a known effect: we have not measured whether such a page makes a practice named more often. And if Gemini tells a patient that a practice nearby is the best choice for nervous patients, what does it say about yours? The audit shows you, question by question.

And your practice, what do the assistants say?

The free Waseit audit asks ChatGPT and Gemini the questions your patients type, in their language and for your city. In about 60 seconds you get a score out of 100, the practices named instead of you, the questions where you do not appear, the tone of the sentences that mention you and, when Gemini searched Google, the pages it actually read. It is free, with no account and no card; an email address only unlocks competitor names and the action plan. It is a dated snapshot, not a ranking: answers change from one day to the next, and we promise no position.

What this does not prove

• These are the models' answers through their programming interfaces, with no user location; ChatGPT and Gemini answered without web search, Perplexity searched by design and was asked in three countries only.
• Each question was asked once, between 30 July and 16 October 2026; ChatGPT here is gpt-4o-mini, not necessarily the model in the ChatGPT app, and the models may have changed since.
• We did not check whether the named practices exist, offer sedation or are gentle with nervous patients: we measure what the assistants say, not whether it is true.
• The count rests on a list of words, read sentence by sentence. A wider list (afraid, worried, phobic) adds 4 ChatGPT answers, all general advice, and 11 Gemini answers, without changing the finding.
• This edition describes assistant answers; it is not medical advice.