Studies / Dentists
We asked AI and Google for the best dentist in two Ontario cities more than 11,500 times. How the question is worded changes what comes back, and Google's AI answer and its map don't agree.
What we found
How people phrase it decides whether Google's AI answer appears at all.
84% vs 1%
Asked as a question, Google showed an AI answer 84% of the time. Typed as keywords, 1% (Peterborough).
Finding 01 ↓A practice can star in Google's AI answer and be missing from its map.
2 in 3 vs 0
One Peterborough practice was in 2 of every 3 Google AI answers, and never once on Google's map.
Finding 02 ↓Ask Claude for "the best" and it usually won't pick one.
9 in 10
Asked "who is the best dentist?", Claude declined to choose 9 times in 10. Asked for "top dentists", it just gave a list.
Finding 07 ↓What the study delivered
The study's job was a number each practice could act on: how often AI and Google name that practice when someone asks for a dentist in its city. Every practice on our list was looked for in every answer, and each practice can get its own private results.
How often it's named
Across all six platforms, as a rate, not a screenshot.
Platform by platform
Where the practice shows up, and where it doesn't.
Against the rest of town
How it compares with every other practice on our list.
And where the answers come from. We also recorded every website the AIs cited along the way: review sites, directories, forums and the practices' own pages. So we can show which sources sit behind each practice's answers, and which ones it's missing from.
Four everyday questions, phrased the way people type them:
Along the way
Measuring a city this thoroughly also shows how the platforms behave. These are the patterns we saw, not the rankings.
01
Typed as keywords ("best dentist in Belleville"), Google showed an AI answer only 1–3% of the time. Asked as a question ("who is the best dentist in Belleville?"), it showed one 84–86% of the time. Capitals, punctuation and "who's" versus "who is" made no difference. Only whether it was a question.
02
Of the five practices each one named most, the AI answer and the map shared only three in Peterborough and two in Belleville. One Peterborough practice was in 2 of every 3 Google AI answers, and never once on Google's map.
3 of 5
top practices shared, Peterborough
2 of 5
top practices shared, Belleville
03
In Peterborough, 4 of the 31 dental businesses on our list were not named once, on any platform, in more than 5,000 answers.
4 of 31
never named, in 5,000+ answers
04
In both cities, no practice was named often without the AI also reading its website. We can't say which comes first: AI may name what it can read, or read what it already names.
05
When the AIs cited a source other than a practice's own site, five or six review and directory websites made up most of it. And each AI has a favourite: ChatGPT leaned on one review site, Claude on others, while Perplexity and Google's AI answer cited practices' own websites most.
74%
of other sources, Peterborough
64%
of other sources, Belleville
06
Claude and Gemini named an individual dentist ("Dr. …") in almost every answer. ChatGPT and Perplexity did so in roughly 60–75% of answers, and more often named the clinic.
07
Asked "who is the best dentist in [city]?", Claude said some version of "there's no single best" in more than 9 answers out of 10. Asked for "top dentists", it almost never did, and simply gave a list.
9 in 10
Claude answers to "who is the best" that declined to pick
Which practices were named, and how often, is in each practice's own report. We don't publish it.
Run a practice in Peterborough or Belleville? You can see how often your practice was named.
The same method works for any question where it matters who gets named.
Claude's figures in this study were collected with an earlier setting that cut off most of its answers. We've since fixed it. Re-checking showed Claude names more practices when its answers aren't cut short, so Claude's figures here are a lower bound.