Blog · 31 October 2026 · 5 min read · By Pogos Kazarian

Case study: 60 ChatGPT answers about dental implants in Tulsa, counted

We asked ChatGPT 12 patient questions about dental implants in Tulsa, five times each, and counted who it named and what it cited. 22 of 51 practices were named; 29 never were.

The short answer: in 60 answers to 12 patient questions about dental implants in Tulsa, Oklahoma, ChatGPT named 22 of the 51 dental practices offering implants that we found in the area at least once. 29 were never named. The leading practice was named in 26 of the 60 answers. The pages behind the answers were mostly practices' own websites, plus healthcare directories, professional associations and review sites, and which practice won depended heavily on the exact question.

What this is and isn't. It is a measurement of one trade in one city, made before we had any clients. There is no "after" yet: no business has been measured with us week after week on Monitor. When one does, and agrees to it, we will publish the before and after here. Practices that come off badly below are not named; practices that ChatGPT named are, because those names are what it said.

The setup

Date: 26 October 2026.
Assistant: ChatGPT's model through OpenAI's API (gpt-5.4), web search required on every answer, approximate location set to Tulsa, Oklahoma.
Questions: 12 that patients ask, from "best dental implant clinic in Tulsa" and "affordable dental implants in Tulsa, Oklahoma" to "I lost a front tooth, where should I get an implant in Tulsa?" and "second opinion on a failing dental implant in Tulsa".
Repeats: each question five times, 60 answers in all. 36 of them ran on one pinned model version and 24 on the model's general name a little earlier; both are gpt-5.4.
Practices: 51 practices offering implants in Tulsa and its suburbs, found by web search per suburb and per specialty. Chains and regional groups were left out.
Counting: a practice counts as named when the answer names it in its text. Every hit was read in full to rule out namesakes.

How we measure describes the method in full.

Who was named

Dental Implant Center of Oklahoma: 26 of 60 answers (43%)
Elite Oral and Maxillofacial Surgery of Oklahoma: 22 of 60 (37%)
Secure Dentures & Implants: 17 of 60
Route 66 Dental Implants & Periodontics: 17 of 60

Below them came a long tail. 22 of the 51 practices were named at least once; 29 were named in none of the 60 answers.

Being named was not one result but twelve. One practice that was named in 3 of 5 answers on both the All-on-4® and same-day implant questions was named in none of the answers to "who is the best implant dentist", "I lost a front tooth" and "board-certified implant dentist". (All-on-4® is a Nobel Biocare trademark.)

What the answers relied on

The two most-cited websites were practices' own: one was linked in 19 of the 60 answers, the other in 18. Both practices were also among the four most named.
A healthcare directory was linked in 13 answers (about 1 in 5).
35 of the 60 answers linked to at least one directory, review site or professional association, such as the American Academy of Implant Dentistry's directory or the American Dental Association's dentist finder.
On the "best implant dentist" and "board-certified" questions, the sources shifted towards credentialing bodies and review aggregators, and away from practices' own pages.

Two pages, compared

Two practices that both offer full-arch implants:

Practice A was named in 0 of 25 answers to five questions where full-arch treatment came up: the best clinic, affordable implants, All-on-4®, same-day implants and full-mouth implants. Its page on the treatment has about 600 words and names the treatment once. It reads as a general page about rebuilding a smile.
Practice B was named in 3 of 5 answers on the All-on-4® question and 3 of 5 on the same-day question, and its page was among the cited sources. Its page on the treatment has about 1,100 words, names the treatment 20 times, says who it suits and answers common questions.

This does not prove that the page is why. It shows what the answers cited, and that Practice A's page gives a reader, human or not, much less to quote.

Suburbs need their own questions

A question about "Tulsa" can't judge a practice in a suburb fairly, so we added three suburb questions, five answers each:

Owasso: Owasso Periodontics & Dental Implants was named in 4 of 5. Three other implant-offering practices in Owasso were named in none.
Jenks and Bixby: Elwood Avenue Dental was named in 5 of 5. Two other practices that offer implants there were named in none.
Glenpool and Sapulpa: 918 Dentist of Glenpool and Eclipse Smiles were each named in 3 of 5.

Each suburb had a clear winner and practices that never appeared, even with the question asked about their own town.

The model matters

Earlier the same day we asked the same 12 questions, five times each, on a smaller, cheaper model (gpt-5.4-mini). The leader changed: on the smaller model a different practice came first. The healthcare directory was cited in 58% of the smaller model's answers and in about a fifth of gpt-5.4's (13 of 60). That is why we pin one model for a series, report it with every number and never compare across models.

What we would tell a practice that was missing

From these answers, not from a general checklist:

One clear page per treatment you want to be recommended for, answering the questions patients ask: what it is, who it suits, how it works at your practice, what it costs if you publish prices. Clinical statements confirmed by the dentist.
The directories and associations the answers cite for your questions: check you are listed, under implants, with the right address.
Reviews where the answers look, asked for honestly and never paid for.
Measure the same questions again after the changes, with enough answers to tell a change from noise.

What this doesn't show

What any real patient saw. Their ChatGPT has its own history and settings; this is a repeatable sample through the API.
Why ChatGPT chose any practice. We report what it said and cited.
Other assistants. This is ChatGPT only.

Want the same count for your business? The free snapshot asks eight questions for your trade and city and shows who is named and what is cited, in about a minute. For why businesses go missing and what to check first, read Why doesn't ChatGPT recommend my business?

ChatGPT is a trademark of OpenAI; Talsara is not affiliated with OpenAI. The named practices are named because ChatGPT named them; we have no relationship with any of them.