Why isn't my dental practice showing up in ChatGPT?
Most practices are missing for a structural reason, not a reputation one. In our June 2026 scan of 212 dental practices across two US cities, ChatGPT named two. The ones it skipped averaged 4.8 stars, so ratings were not the blocker. The measurable gaps came earlier: whether an engine can reach the site, and whether anything on it describes the practice in machine-readable form. Only 6 of 212 had that.
If you have asked ChatGPT for the best dentist in your city and watched it name four practices that are not yours, you already know the feeling. The useful question is which part of the chain broke, because the fix for a crawler problem has nothing in common with the fix for a mentions problem.
We can answer that with our own numbers rather than opinion. In June 2026 we scanned 154 dental practices in Charlotte, NC and 58 in Columbus, OH. For each one we recorded what ChatGPT and Perplexity said when asked patient-style questions, plus the site's crawler access, structured data, Bing footprint and review profile. One Columbus batch went out with a malformed city label in the query, and the Columbus report spells out what that limits. The full method is written out in how we measure AI visibility.
| What we checked | Charlotte (154 practices) | Columbus (58 practices) |
|---|---|---|
| Named by ChatGPT | 2 | 0 |
| Named by Perplexity | 3 of the 135 we could read | 3 of the 21 we could read |
| Average rating of the practices ChatGPT skipped | 4.82 stars | 4.74 stars |
| Practices at a perfect 5.0 that ChatGPT skipped | 41 | 15 |
| Full structured data in place | 3 of 154 (2%) | 3 of 58 (5%) |
| robots.txt open to AI crawlers | 147 of 154 (95%) | 54 of 58 (93%) |
| Gemini | Not included in this scan | Not included in this scan |
That floor is not unique to two cities. In its 2026 Local Visibility Index, covering roughly 350,000 locations across 2,751 brands, SOCi measured how often local businesses get recommended by each assistant.
| Surface | Locations recommended nationally | Our June 2026 dental scan |
|---|---|---|
| ChatGPT | 1.2% | 2 of 212 |
| Perplexity | 7.4% | 6 of the 156 we could read |
| Gemini | 11% | Not measured |
| Google local 3-pack | 35.9% | Not measured |
Our Charlotte rate, 2 practices out of 154, lands on the national ChatGPT figure almost exactly. That is a sign the scan is reading reality rather than a local quirk.
Are my reviews the reason ChatGPT ignores my practice?
Almost certainly not. The practices ChatGPT skipped in our June 2026 scan averaged 4.82 stars in Charlotte and 4.74 in Columbus. Forty-one Charlotte practices and 15 in Columbus sat at a perfect 5.0 and were still not named. One invisible practice had 2,972 reviews.
Reputation is the first thing owners suspect, and it is the thing our data clears fastest. In Charlotte, 138 of the 151 skipped practices with rating data were at 4.5 stars or better, and the median review count among them was 264.
Set that against the national picture. Locations ChatGPT recommended averaged 4.3 stars in SOCi's 2026 index. The practices our scan found invisible were already sitting above it. Ratings work as a filter on candidates an engine has found, not as a way to find new ones, which is why clearing that level does not get you named.
In Charlotte, a practice with a 5.0 rating and nearly 3,000 reviews was named by neither ChatGPT nor Perplexity. It also had no structured data on its site. Full city numbers are in the AI visibility report for Charlotte dentists.
Where does ChatGPT look when a patient asks for a dentist?
Not Google. For local questions ChatGPT runs a Bing search and works from those results. "ChatGPT performs a Bing search for the relevant local query, gathers 20-30 of the top web results returned by Bing, and organizes those results using its own logic," wrote Damian Rollison in Search Engine Land in 2025. If Bing holds little about you, the assistant has little to pick from.
This is why practices with respectable Google rankings can still be missing. The two indexes are not the same, and dental sites are usually built and marketed with only Google in mind. In Charlotte we flagged at least 22 of the 154 sites as having a thin Bing footprint, roughly 19 indexed pages or fewer, and 13 more could not be read at all.
Being findable in Bing is the entry condition, not the promise: it puts you in the pile the assistant reads, and nothing more. Step-by-step work on the Bing side is covered in our guide to getting your practice recommended by ChatGPT.
Source: Damian Rollison, Search Engine Land, May 2025; Citevio scan, June 2026.
My site is in Bing's index. Why am I still not named?
Because the assistant picks a handful of names out of 20-30 results and has to stand behind each one. In Columbus we could read Bing coverage for 21 of the 58 practices. All 21 were in the index, from about 50 pages to about 145,000. ChatGPT named none of them.
Indexing gets your pages into the pile an engine reads. Selection is a separate contest, decided by how confidently the engine can describe you and how often it has seen you confirmed somewhere it did not get from you.
One limit worth stating on that Columbus finding: those 21 practices come from the batch of our scan where the city label inside the query was malformed, so treat the number as directional. Which checks failed, and why, is written into the AI visibility report for Columbus dentists. The Charlotte side points the same way from a cleaner angle: 147 practices had their site fully open to AI crawlers, and ChatGPT named 2 of them.
Can an AI engine tell what my practice actually does?
Often it cannot. In Charlotte, 3 of 154 practices had complete structured data on their site and 151 had none or only fragments. Columbus was close to the same, 3 of 58. Structured data is how a page states its own facts in machine-readable form: what the business is, where it is, what it treats.
One caution, because this area attracts bad advice. Structured data is plumbing, not a citation switch. In 2026 Mark Williams-Cook hid a made-up address inside deliberately invalid schema markup on a test page, with nothing in the visible copy, and both ChatGPT and Perplexity repeated the address back. His reading: "they were not parsing it as schema. They were doing what LLMs always do: reading the visible-ish text of the page."
What gets quoted is text that answers a real patient question in full, in one place, without the reader having to assemble it from three sections. So the same fix serves both purposes: write the answer to "how much does Invisalign cost in Charlotte" as a clear block of text on the page, then let the structured data describe it.
Source: Mark Williams-Cook, 2026; Citevio scan, June 2026.
How much do mentions on other sites matter?
More than anything you control on your own site. Across 75,000 brands, Ahrefs found that mentions of a brand on other websites correlated with appearing in Google's AI Overviews at 0.664, while links pointing at the brand's site sat at 0.218. Brands in the top quarter for mentions averaged 169 AI Overview appearances, against 14 for the next quartile down.
That study looked at Google's AI Overviews rather than ChatGPT, so treat it as evidence about how these systems weigh outside confirmation, not as a ChatGPT-specific rule. The pattern repeats elsewhere. Muck Rack analyzed more than 25 million links cited by ChatGPT, Claude and Gemini and found 84% came from earned media, meaning coverage and third-party sources rather than a company's own or paid content. Paid and advertorial content accounted for 0.3%.
For a dental practice that usually means directories, local press, dental association listings, patient-community threads and review platforms carrying a consistent name, address and phone number. A practice that exists only on its own website is asking an assistant to vouch for a business nobody else has confirmed.
Sources: Ahrefs, May 2025; Muck Rack, May 2026.
Is my website blocking the AI crawlers?
Probably not, and it is worth ruling out in one minute. In our June 2026 scan, 4 of 154 Charlotte practices and 2 of 58 in Columbus were blocking AI crawlers in robots.txt. It is the rarest reason we found and the only one that is absolute: a crawler that cannot read the page has nothing to quote.
Blocks usually arrive by accident, through a plugin default or a developer's staging rule that nobody removed. Open yoursite.com/robots.txt and look for lines naming GPTBot, OAI-SearchBot, PerplexityBot or Google-Extended with a Disallow rule underneath. We measured how common this is nationally: of 6,497 US dental practice websites, 11.4% block at least one of the crawlers we test, and 7.6% block one run by OpenAI, Anthropic or Perplexity. The per-crawler and per-state breakdown is in are dental websites blocking AI crawlers?
Then set expectations honestly, because the same data cuts the other way: 145 of the 147 Charlotte practices with a fully open robots.txt were still not named by ChatGPT. Opening the door is necessary. It is not persuasive. An open robots.txt is also not the last check: among sites whose robots.txt already welcomes AI crawlers, 14.4% had the request refused by their own server anyway, and 12.6% served Googlebot normally while refusing an AI crawler. That gap is measured in open to AI on paper, blocked in practice.
What if the practice is new and there is almost nothing to read?
Then the reason is usually different from every other reason on this page. A new practice is generally not being blocked or filtered out — there is simply not yet a record for these systems to compile. Nothing has been crawled, no third-party data lists you, nobody has contributed anything about you. The first job is not ranking above anyone. It is existing in a readable form.
That reframes what to do first. Being invisible because a server refuses a crawler is a fault to fix; being invisible because you opened in March is an absence to fill, and the two need different work. For a new practice the sequence is: a site whose content is in plain text and reachable, a claimed and complete Google Business Profile, and consistent details wherever your name already appears — before anything that could be called optimisation.
It also helps to know the distance honestly. Across 523 practices we could read review data for, the median sat at 250 reviews and a 4.9 median rating. A practice in its first year starts a long way from that distribution. That is not a target to hit and not a bar an engine is known to apply — it is an expectation setting, so that six months of steady work does not get mistaken for failure against a number nobody told you about.
Source: Citevio dental practice web anatomy study, 6–26 June 2026. Review figures n = 523 of 527 practices.
One measured difference by specialty is worth knowing, with a strict caveat attached. In the July 2026 robots.txt study, orthodontic sites blocked AI crawlers noticeably less often than general dental ones: 4.8% of orthodontist domains (n = 415) blocked at least one of thirteen AI and data crawlers, against 12.5% of dentist domains (n = 3,929). That is a difference in what permission files declare — it is not evidence that orthodontists get named more often, and no such measurement exists here.
Source: Citevio AI crawler access study, 22 July 2026, n = 6,497 readable domains. Specialty figures describe crawler blocking declared in robots.txt, not AI visibility.
Both the crawler and web-anatomy evidence cited in this diagnosis are deposited under DOI 10.5281/zenodo.22016876.
The question every new practice asks next is how long this takes, and we do not have an answer to sell you. No timeline for a new practice appearing in an AI answer has been measured here, so the honest response is a check schedule rather than a date — set out in how long this takes, and why we will not give you a date. On budget, we hold no dental spend or ROI data at all, so there is no "new practice should spend X" figure on this site, only a difference in kind: at this stage the money goes into being readable and being listed, not into competing for a position.
The checks themselves cost nothing but attention and are written out on how to test if AI can read your dental website. Where the review distribution puts you, and what a compliant way of asking looks like, is on how many Google reviews a dental practice has.
So which reason applies to your practice?
Crawler blocks, a thin Bing footprint and missing structured data all leave a trace on your own site, and the checker below reads your site and scores its AI-readiness in seconds. Mentions are the one you check by hand: search your practice name and see who besides you confirms it exists. If you would rather run the engine tests yourself, our walkthrough on how to check what ChatGPT says about your practice lists the exact queries to use.
Prefer a deeper look? The full AI visibility checker for dentists tests what ChatGPT, Perplexity, Gemini and Google AI Overviews actually say about you.
One caveat applies to any snapshot, including ours. Profound tested about 80,000 prompts per platform and found that roughly 40-60% of the domains cited in AI answers were completely different one month later, for identical questions. A practice that was invisible in June may surface in August, and the reverse. That is why every number on this page carries a scan date, and why we re-scan instead of quoting a figure forever.
Source: Profound, July 2025.
Citevio's dental market research is published openly under CC-BY-4.0 with a DOI, so the numbers can be checked and reused; client data is never published.
Common questions
Why does ChatGPT recommend my competitor instead of me?
Usually because the assistant has more to go on for them. It builds local answers from a Bing search plus whatever else it can verify, so a competitor who is indexed, described clearly on their own site, and mentioned on sites they do not own is easier to name with confidence. In our June 2026 scan of 212 practices, rating gaps did not explain who got named. Visibility gaps did.
Do I need to be on page 1 of Google to show up in AI answers?
Not for ChatGPT. It runs a Bing search for local questions and gathers 20 to 30 of the top results from there, according to Search Engine Land's 2025 breakdown, so your Google position is not the gate. Google's AI Overviews are a separate surface and still lean on Google's own results. Being findable in Bing is the part most dental sites overlook.
Does blocking AI bots in robots.txt hurt my practice's visibility?
Yes, and the effect is absolute: an engine that cannot read your site has nothing to quote. It is also rare. In our June 2026 scan, 4 of 154 Charlotte practices and 2 of 58 in Columbus blocked AI crawlers outright. The catch is that opening the door does not get you named. Of the 147 Charlotte practices with an open robots.txt, ChatGPT named 2.
What percentage of dental practices did ChatGPT actually name in your scan?
Two out of 212. We scanned 154 practices in Charlotte and 58 in Columbus in June 2026; ChatGPT named 2 in Charlotte and 0 in Columbus. Our Charlotte rate lands almost exactly on the 1.2% SOCi measured nationally across all local businesses, not just dental practices.
Did Perplexity name more practices than ChatGPT in your scan?
A few more, but still a small share. Perplexity named 3 of the 135 Charlotte practices we could get a reading for, and 3 of the 21 readable in Columbus. Some checks failed to return a reading and were dropped from that row rather than counted as invisible.
Was Gemini included in this scan?
No. Our June 2026 scan of 212 practices covered ChatGPT and Perplexity only; Gemini was not measured in this study.
How many perfect 5-star practices did ChatGPT still skip?
Forty-one in Charlotte and 15 in Columbus. Practices at a flawless rating were passed over just as often as lower-rated ones, which is part of why we rule out reputation as the main blocker.
What is the biggest structural gap you found between named and skipped practices?
Structured data. Only 3 of 154 Charlotte practices and 3 of 58 in Columbus had complete, machine-readable structured data describing what the practice is and does. That gap was far larger than any difference in ratings or review counts.
Do practices with more reviews get named more often?
Not in what we measured. One Charlotte practice with a 5.0 rating and nearly 3,000 reviews was named by neither ChatGPT nor Perplexity, and it also had no structured data on its site. Review volume did not separate the named practices from the skipped ones.
How does the median skipped practice compare to what SOCi found nationally?
It is already above the bar. Locations ChatGPT recommended nationally averaged 4.3 stars in SOCi's 2026 index; in Charlotte, 138 of 151 skipped practices with rating data were at 4.5 stars or better, with a median of 264 reviews. Clearing the reputation bar did not get them named.
Why does ChatGPT care about Bing instead of Google?
Because that is where it looks. For local questions, ChatGPT runs a Bing search and works from those results — Search Engine Land described it in 2025 as gathering 20 to 30 of the top Bing results and organizing them with its own logic. A site with little presence in Bing has little for the assistant to draw on, regardless of its Google ranking.
How common is a thin Bing footprint among dental practices?
Common enough to matter. In Charlotte we flagged at least 22 of 154 sites with roughly 19 indexed Bing pages or fewer, and 13 more that could not be read at all. Most dental sites are built and marketed with only Google in mind.
If my site is fully indexed in Bing, will ChatGPT name me?
Not necessarily. In Columbus, all 21 practices we could check were indexed in Bing, with page counts from about 50 to roughly 145,000, and ChatGPT named none of them. Indexing gets you into the pile the assistant reads; selection is a separate contest.
Does an open robots.txt mean my practice will get named?
No, it only removes one blocker. In Charlotte, 147 of 154 practices had a fully open robots.txt, and ChatGPT still only named 2 of them. Opening the door is necessary; it is not persuasive on its own.
How many practices in your scan were actively blocking AI crawlers?
Very few: 4 of 154 in Charlotte and 2 of 58 in Columbus. It is the rarest cause we measured, but also the only one with an absolute effect, since a crawler that cannot read the page has nothing to quote. The national picture across all US dental sites is set out on the page about whether dental websites are blocking AI crawlers.
Can my site pass every check on this page and still not get named?
Yes. In Charlotte, 145 of the 147 practices with a fully open robots.txt were not named by ChatGPT. An open robots.txt, indexed Bing pages and good reviews clear the entry conditions; they do not guarantee a mention.
Why did you flag invalid schema markup as a caution rather than a fix?
Because a 2026 test by Mark Williams-Cook showed both ChatGPT and Perplexity repeated a made-up address that was hidden inside deliberately invalid schema, with nothing in the visible copy, meaning they were reading the page's visible-ish text rather than parsing the schema itself. Structured data should describe an answer that is also written out in plain text, not substitute for it.
Do mentions on other websites matter more than what is on my own site?
Based on the research we cite, yes. Across 75,000 brands, Ahrefs found mentions of a brand on other sites correlated with AI Overview visibility at 0.664, against 0.218 for links pointing at the brand's own site. Muck Rack separately found 84% of the links ChatGPT, Claude and Gemini cite came from earned media, not owned or paid content.
My practice is brand new. Does all of this still apply?
Not in the same way. A new practice is usually not being filtered out; there is simply not yet a record for these systems to compile, since nothing has been crawled, no third-party source lists it, and nobody has confirmed it elsewhere. The first job is existing in readable form, not competing for a position, and we do not have a measured timeline for how long that takes.
How confident are you that these findings hold beyond two cities?
Reasonably confident on the ChatGPT number specifically. Our Charlotte rate of 2 named out of 154 lands almost exactly on the 1.2% SOCi measured nationally across roughly 350,000 locations, which suggests the scan is reading a real pattern rather than a local quirk. The Columbus figures carry a caveat: that batch had a malformed city label in the query, so treat those numbers as directional.