Ask an AI engine for a dentist in a US city and it answers with a short list of names, then keeps returning the same ones. Across 580 runs in seven metros in June 2026, the most-named practice in each metro appeared in at least half of the runs that produced any name, and in three metros in more than nine out of ten. Every city report is linked below.
How many dental practices does AI name in one city?
Fewer than most owners expect, and the same ones repeatedly. In our June 2026 scans, a single run of a city's question set returned a median of 6 to 20.5 distinct practice names, depending on the metro. Half of every name appearance in a metro came from between 11 and 29 practices. In Austin, the top name appeared in 133 of the 134 runs that produced any name.
The table below has not been published anywhere before. Each row is one metro. Read the column headings carefully, because two of them count different things: a run is one pass of that city's four patient-style questions, and there were many runs per metro. A run is not a practice, and the totals here are not the practice counts used in the individual city reports.
| Metroscan window, June 2026 | Runspasses of that city's 4 questions | Runs naming at least one practiceout of the runs in this row | Distinct practice names producedacross all runs in this metro | Median names per naming rundenominator: naming runs | Top name's appearance rateshare of naming runs | Practices covering half of all name appearancesfewer means more concentrated |
|---|---|---|---|---|---|---|
| Charlotte, NCJun 6–23 | 177 | 160 | 338 | 19 | 83.1%133 of 160 | 12 |
| Austin, TXJun 24–25 | 134 | 134 | 243 | 20 | 99.3%133 of 134 · three-way tie | 15 |
| Raleigh, NCJun 24 | 87 | 86 | 269 | 7 | 94.2%81 of 86 | 26 |
| Nashville, TNJun 23 | 68 | 68 | 178 | 6.5 | 64.7%44 of 68 | 16 |
| Tampa, FLJun 24 | 51 | 51 | 172 | 6 | 68.6%35 of 51 | 21 |
| Columbus, OHJun 23 | 43 | 41 | 155 | 7 | 51.2%21 of 41 | 29 |
| Salt Lake City, UTJun 25–26 | 20 | 20 | 54 | 20.5 | 100.0%20 of 20 · three-way tie | 11 |
| All seven metros | 580 | 560 | see rows | — | — | — |
Source: Citevio mention-concentration analysis, built from 580 dental scan records across 7 metros, June 6–26, 2026. Engines queried: ChatGPT and Perplexity. Gemini was not included in this scan. Every figure in this table was regenerated from the raw scan files on 2 August 2026 before publication. Metros with fewer than 20 runs are excluded; Salt Lake City sits exactly at the 20-run floor and its row should be read as the thinnest in the table.
Two rows deserve a second look. Austin and Salt Lake City show a tie at the top: three different practices each appeared in the same number of runs, so neither metro has a single leader. And the spread across metros is wide. In Charlotte, 12 practices accounted for half of every name appearance across 160 naming runs. In Columbus, it took 29. For national context, SOCi's 2026 Local Visibility Index found that ChatGPT recommended just 1.2% of the locations it studied, against 7.4% for Perplexity and 11% for Gemini, and concluded that "AI visibility is three to 30 times harder to achieve than ranking well in traditional local search."
National comparison: SOCi 2026 Local Visibility Index, reported by Search Engine Land, 28 January 2026, based on nearly 350,000 locations across 2,751 multi-location brands. searchengineland.com
What do these concentration numbers mean, and what don't they?
They mean the list AI hands a patient is short and repetitive, not that any practice earned its place. This data records which names came back, never why. A practice may be named often because of its reviews, its Bing footprint, how often other sites mention it, or something we never measured. These are associations, not causes, and nothing on this page should be read as a ranking factor.
Three things the table does say, stated plainly:
- The answer set is short. A patient does not get ten links to sort through. In four of the seven metros, a typical naming run produced 7 or fewer distinct practice names.
- The same names repeat. In every metro we scanned, one practice appeared in at least half the naming runs, and in Austin and Raleigh in more than nine out of ten.
- How crowded that list is varies by city. Charlotte and Salt Lake City concentrated half of all name appearances into 12 and 11 practices. Columbus spread the same half across 29.
What we cannot tell you is the mechanism. The best available outside evidence points the same way but stops short of causation too: Ahrefs studied 75,000 brands in May 2025 and found brand web mentions correlated with AI Overview mentions at 0.664, while backlinks managed 0.218 and referring domains 0.295. The authors put the caveat in their own words: "While the data shows statistical relationships, I should emphasize that correlation ≠ causation." We apply the same caution to our own numbers.
Correlation figures: Louise Linehan and Xibeijia Guan, Ahrefs, 26 May 2025, 75,000 brands. ahrefs.com
Which city report should you read?
Read the one for your market. Six metros have a full report of their own with engine-by-engine visibility rates, and the national master report pulls all seven together. Each city report answers a different version of the same question: out of the practices we scanned in that city, how many did ChatGPT and Perplexity actually name, and what did the ones they skipped have in common?
The practice counts below come from each report and are the hand-checked figures those pages publish. They are smaller than the run counts in the table above, because a report counts verified dental practices while a run counts one pass of the city's questions.
| Report | Practices scannedhand-checked, per that report | What that report tells you |
|---|---|---|
| The 2026 Dental AI Visibility Reportnational master, all 7 metros | 527 | Every metro in one place, plus a query-intent layer showing that asking for a cosmetic, Invisalign or implant dentist barely changed the result, and what the 527 sites look like on the inside. |
| AI visibility report: Charlotte dentists | 154 | ChatGPT named 2 of 154 practices (1.3%); Perplexity named 3 of the 135 it returned a readable result for. The highest full-sample ChatGPT rate of the six cities. |
| AI visibility report: Austin dentists | 111 | Perplexity named 7 of 111 practices (6.3%), the strongest Perplexity result we recorded. ChatGPT hit a rate limit on 90 of them, so its result rests on the 21 we could read. |
| AI visibility report: Raleigh dentists | 64 | ChatGPT named zero of 64 practices on a fully readable base. Perplexity is not scored there, because 63 of 64 queries returned an authentication error rather than a result. |
| AI visibility report: Nashville dentists | 59 | ChatGPT named zero of 59 practices on a fully readable base. Every Perplexity query returned an authentication error, so no Perplexity rate is published for that city. |
| AI visibility report: Columbus dentists | 58 | ChatGPT named zero of 58 practices; Perplexity named 3 of the 21 it returned a readable result for. The least concentrated metro in the table above. |
| AI visibility report: Tampa dentists | 42 | ChatGPT named zero of 42 practices. Perplexity could not be read at all in that scan, so Tampa has a ChatGPT result and no Perplexity result. |
Figures copied unchanged from each published report. If any number differs elsewhere on this site, the city report is the one to trust. Salt Lake City has no report of its own; its practices sit inside the 527-practice national total.
What this measurement does not tell you
This is one measurement, taken in a June 2026 window. We have not yet measured whether AI visibility is getting harder or easier over time, and we will not imply it. When we have a second scan, we will publish the difference, including any part of it that does not favor us. Until then, treat every figure here as June's weather.
The honest limits of this dataset, all of them:
- No trend, in either direction. One scan window produces a snapshot and nothing else. Anyone selling you a trend from a single measurement is selling you an opinion with a number attached.
- Answers move on their own. Profound tested roughly 80,000 prompts per platform in July 2025 and found that "roughly 40-60% of the domains cited in AI responses will be completely different just one month later." A June result is not a permanent verdict.
- Runs are not practices. The run counts here include records that the hand-checked city reports removed, such as duplicate listings and businesses that turned out not to be dental practices at all. Never read a run count as a practice count.
- Two engines, not three. These runs queried ChatGPT and Perplexity. Google Gemini was not included in this scan, and we publish no Gemini figure of our own. In some metros Perplexity returned authentication errors, so the names collected there came from ChatGPT alone. We have not split these results by engine.
- Names were read by pattern matching. Practice names were extracted from answer text automatically, then filtered to remove sentence fragments, generic phrases and institutions. A misread name can survive that filter, which is why we publish counts rather than leaderboards.
- One batch was thrown out. Twenty-one Columbus runs carried a corrupted city label that went into the query text itself, so the engines were asked a question we never intended. Those runs are excluded here, exactly as they are excluded from the national report, rather than quietly folded in.
- The population is metros, not the country. Seven metros are seven metros. Nothing here supports a sentence beginning "US dental practices."
Volatility figure: Profound domain-drift analysis, 17 July 2025, approximately 80,000 prompts per platform. tryprofound.com
How did we measure this?
For each practice in a metro, we ran that city's four patient-style questions through ChatGPT and Perplexity, recorded whether the practice itself was named, and captured the other practice names the answer contained. Repeating that across a metro turns a single answer into a distribution: which names come back, how often, and how many practices it takes to cover half of all appearances.
The questions were the ones a patient would type, with the city name filled in: best dentist, best cosmetic dentist, best Invisalign dentist, best dental implant dentist. That wording matters for a reason worth stating clearly. The practices we scanned were not selected for being cosmetic clinics. The label belongs to the question, not to the practice, so no figure on this page describes "cosmetic dentists" as a group. Most Charlotte practices were tested on those four questions and an earlier batch of 28 runs used a slightly different local set, which the Charlotte report documents in full.
Every number in the table above was regenerated from the raw scan files on 2 August 2026, before this page was written, using a script kept alongside the data. Doing that caught two errors in our own stored figures and both were corrected before publication rather than footnoted. The full method sits in how we measure AI visibility, and the underlying scan data, including downloadable CSVs, is on our open data page.