We scanned 527 dental practice websites across seven US metros in June 2026 and asked ChatGPT and Perplexity the four ways a patient actually phrases a search: general, cosmetic, Invisalign and implant intent. The query wording barely mattered. ChatGPT named a practice on 0.0% to 0.3% of the readings we completed, across all four intents; Perplexity, on 1.6% to 2.4%.
This page pulls together every Citevio dental scan into one report: the six metros with their own full city report, the site-level anatomy of what these 527 practices are built with, and a query-intent layer that has not appeared on this site before. It does not repeat any single city's findings in full; each metro links out to its own report for the detail.
How was this scan run?
Between 6 and 26 June 2026 we scanned 527 dental practice websites in seven metros, deduplicated to one row per domain. We asked ChatGPT and Perplexity patient-style questions for each practice and recorded whether it was named. Not every query returned a usable answer: rate limits and authentication errors are common with both APIs, and a failed call is recorded as a gap, never as "not recommended."
Two different things get measured on this page, and they use different denominators, which is why no single "N" applies to every number below. The city reports ask each engine once per practice for a general "best dentist" query and report the practice-level outcome. The query-intent layer, new in this report, asks four separate questions per practice, general, cosmetic, Invisalign and dental implant intent, and scores each one on its own, so a practice can have a valid reading for one intent and an error for another.
| What we tried to measure | Result |
|---|---|
| Practices scanned (deduplicated by domain) | 527, across 7 metros plus 9 smaller towns |
| Practices with at least one usable query-intent reading | 506 of 527 (96.0%) |
| Of those, dropped for a corrupted city label in the query text | 20, see note below |
| Practices left in the published query-intent table | 486 of 527 (92.2%) |
One batch of scans stored "Columbus" under a corrupted city label, "Columbus Oh Cosmetic Dental Clinics", left over from the source list we scanned from. For those 20 practices, the text actually sent to each engine read "best dentist in Columbus Oh Cosmetic Dental Clinics" instead of "best dentist in Columbus," so the reading does not represent the question we intended to ask. We are not quietly folding those readings into a query type they do not represent: all 20 are excluded from the query-intent table below, on every engine and every query type, and we are stating why here rather than smoothing it out of the numbers.
Source: Citevio dental practice web anatomy study, June 2026 (n = 527), and the query-intent layer built on the same 527-practice set. Full method: how we measure AI visibility.
How does each metro compare?
Every metro sits near the floor, and no city stands out as clearly better or worse once you account for how much of each engine we could actually read. Charlotte's ChatGPT rate, 2 of 154, is the highest full-sample ChatGPT result of the six; three metros, Raleigh, Nashville and Tampa, had ChatGPT name zero practices out of a fully readable base.
| Metro | Practices scanned | ChatGPT named | Perplexity named | Full report |
|---|---|---|---|---|
| Charlotte, NC | 154 | 2 of 154 (1.3%) | 3 of 135 readable (2.2%) | Charlotte report |
| Austin, TX | 111 | 1 of 21 readable (90 hit a rate limit) | 7 of 111 (6.3%) | Austin report |
| Raleigh, NC | 64 | 0 of 64 (fully readable) | Not scored (63 of 64 queries hit an authentication error) | Raleigh report |
| Nashville, TN | 59 | 0 of 59 (fully readable) | Not scored (every query hit an authentication error) | Nashville report |
| Columbus, OH | 58 | 0 of 58 (fully readable) | 3 of 21 readable | Columbus report |
| Tampa, FL | 42 | 0 of 42 (fully readable) | Not scored (Perplexity was not readable in this scan) | Tampa report |
These six numbers are copied unchanged from each metro's own published report; if you find a different figure anywhere else on this site, the city report is the one to trust. A seventh metro, Salt Lake City, and 21 practices across nine smaller towns near these six cities are part of the 527-practice scan behind the rest of this page, but their samples are too thin to score city by city, so they are not broken out here. Practice counts for the same metro can differ by one or two between this dataset and a hand-checked city report; the data page explains why in detail.
Does the way a patient asks change the answer?
Not by much. We split the same 527-practice scan into four query intents, general, cosmetic, Invisalign and dental implant, and scored each one separately, after dropping 20 readings whose query text carried a corrupted city label. ChatGPT stayed at 0.0% to 0.3% across all four; Perplexity ranged from 1.6% to 2.4%. Neither engine got meaningfully more generous when the question named a specialty.
Read this table carefully: the denominator changes row by row, because each query was scored independently and not every call to ChatGPT or Perplexity returned a usable answer. A practice missing from one row's denominator simply has no reading for that specific query, not a "not recommended" result. And these practices were not pre-sorted by specialty; the 372 in the cosmetic row are the dental practices we had a valid reading for on the cosmetic-intent query, not 372 practices that identify as cosmetic dentistry clinics.
This table and the city table above count different things, so their figures are not meant to add up. A city row counts a practice once if any of the four queries named it, which is why Austin shows 7 named practices for Perplexity while no single query row here reaches that number across all metros combined. A row here counts one reading at a time, and the same practice can appear in more than one row.
| Engine | Query asked | Valid readings | Named the practice | Share |
|---|---|---|---|---|
| ChatGPT | "best dentist in {city}" | 373 | 1 | 0.3% |
| ChatGPT | "best cosmetic dentist in {city}" | 372 | 0 | 0.0% |
| ChatGPT | "best Invisalign dentist in {city}" | 371 | 1 | 0.3% |
| ChatGPT | "best dental implant dentist in {city}" | 372 | 1 | 0.3% |
| Perplexity | "best dentist in {city}" | 245 | 6 | 2.4% |
| Perplexity | "best cosmetic dentist in {city}" | 244 | 4 | 1.6% |
| Perplexity | "best Invisalign dentist in {city}" | 245 | 4 | 1.6% |
| Perplexity | "best dental implant dentist in {city}" | 245 | 4 | 1.6% |
Source: Citevio query-intent layer, built on the June 2026 practice web anatomy scan (n = 527), reproduced independently from the underlying scan files for this report; corrupted-city-label readings excluded rather than reclassified. Playbook note on scope: clinics were not labeled "cosmetic" in our data; the query sent to each engine was.
An Invisalign-intent query barely moving ChatGPT's or Perplexity's answer does not mean adding an Invisalign page to your site is pointless. We did not test that. What we measured is what these two engines did with a specific phrase in June 2026, on the practices we had a reading for. A practice's own content, mentions elsewhere and crawl history all factor into whether an engine can find it in the first place, and none of that is captured by the query wording alone.
For scale, a 2026 study of roughly 350,000 business locations across categories put ChatGPT's local visibility rate at 1.2% and Perplexity's at 7.4%. Our general-intent row lands close to that ChatGPT figure (0.3%, on a much smaller and single-vertical sample) and below it for Perplexity (2.4%), which is one more reason we read this as "the floor is low everywhere for a single-location dental practice" rather than a dental-specific problem.
Source: SOCi 2026 Local Visibility Index (~350,000 locations, 2,751 brands), reported by Search Engine Land, January 2026; measured nationally across business categories, not dental-specific, so read the two figures side by side rather than as a like-for-like gap.
What do these 527 websites look like on the inside?
Reputation is not the gap. 96.7% of the practices with a Google rating already sit at 4.3 stars or higher. The gap is machine-readable infrastructure: 53.9% of the 527 sites carry no LocalBusiness or Dentist schema at all, and only 3.4% have both LocalBusiness and FAQPage schema in place.
| Signal | Finding | Detail |
|---|---|---|
| Structured data | 53.9% have no LocalBusiness/Dentist schema at all | why AI recommends some dentists |
| robots.txt | 93.4% open to the AI crawlers we tested | are dental sites blocking AI crawlers |
| Site speed | Median First Contentful Paint 3.01s (n=256 of 527 completed a reading) | how slow are dental websites |
| Publishing platform | 54.8% WordPress (289 of 527) | what platform do dental sites run on |
| Google rating | Median 4.9 stars; 96.7% at 4.3 or higher (n=523) | is it their reviews? |
The 3.01-second median First Contentful Paint sits just past the 3.0-second line Google's own web.dev guidance uses to mark a page "Poor," and the robots.txt figure describes what 527 sites' rules literally say, not what their servers do when asked; a missing robots.txt file is a gap in control, not a block, under the current Robots Exclusion Protocol standard. None of this proves that fixing schema or speed makes an engine name a practice. What it shows is what the practices AI almost never names have in common, which is the closest thing to a lead this kind of data can honestly offer. The full argument, including a controlled test that found engines reading page text rather than schema, is in why AI recommends some dentists and not others.
Sources: web.dev, "First Contentful Paint (FCP)", Google, checked 26 July 2026; RFC 9309, Robots Exclusion Protocol, IETF, 2022.
What this report does not claim
Six limits, stated plainly. This is not a national number, Gemini is not in this scan, one snapshot is not a trend, and nothing here proves that any single fix changes whether an engine names a practice.
- Not a national statistic. These 527 practices sit in seven metros plus nine smaller towns. "US dental practices" as a phrase belongs only to our separate 6,497-domain robots.txt study, not to this practice-level scan.
- Gemini is not included in this scan. We did not run it through the same pipeline as ChatGPT and Perplexity here. That is a measurement gap, not a finding of zero visibility, and we are not going to write "0%" for an engine we did not test.
- One time point, not a trend. Every number on this page carries the June 2026 scan window. We do not yet have a second scan of the same practices to compare against, so nothing here describes whether visibility is improving or getting worse.
- Description, not cause. The schema, robots.txt and review figures describe what the practices AI rarely names have in common. We have not run a controlled test that isolates any one of them as the reason, and we are not claiming one does.
- The query-intent layer covers two engines. ChatGPT and Perplexity only; it does not extend to Gemini or to any AI Overview-style result.
- A city report and this report can differ by a practice or two for the same metro. They were filtered separately; where they differ, the standalone city report is the number to cite.
Get the underlying data
Every figure on this page comes from a dataset that is free to download as CSV, with scan dates, denominators and what we chose not to publish. Read citevio.com/data for the full library, or see your own site read against the same categories below.
See what AI says about your practice, free, 2 minutes.
Common questions
How many dental practices does this report cover?
527 practice websites across seven US metros, Charlotte, Austin, Raleigh, Nashville, Columbus, Tampa and Salt Lake City, plus 21 practices in nine smaller nearby towns, scanned in June 2026. Six of those metros also have their own full city report, linked above; Salt Lake City and the smaller towns are too thin a sample to score on their own, so they appear only inside the 527-practice total.
Does asking for a cosmetic or Invisalign dentist change whether AI names a practice?
Barely. Across the four ways we asked, general, cosmetic, Invisalign and implant intent, ChatGPT named a practice on 0.0% to 0.3% of the readings we completed, and Perplexity on 1.6% to 2.4%. The practices we scanned were not sorted by specialty; every percentage is out of the full set of dental practices we had a reading for on that specific query, not a subset of cosmetic-only clinics.
Why isn't Gemini included in this report?
We did not run Gemini through the same automated pipeline as ChatGPT and Perplexity for this scan, so it is not included here. That is a gap in what we measured, not a finding that Gemini shows zero practices. We are not willing to publish a number for an engine we did not test.
Is this a national statistic for US dental practices?
No. This report covers 527 practices in seven metros plus nine smaller towns, not a national sample. The only figure on this site that is national in scope is our separate robots.txt study of 6,497 dental domains; the practice-level findings here describe these specific metros in June 2026.
Where can I get the underlying data?
Every dataset behind this report is a free CSV download at citevio.com/data, including scan windows, denominators and what we chose not to publish. The city reports linked in this report each have their own page.