In our June 2026 scan of 58 Columbus, OH dental practices, ChatGPT named zero of them when a patient asked for the best dentist in town. Not one. Perplexity named 3 of the 21 practices it returned a readable result for. Below are the queries we ran, what the invisible practices share, and why being indexed and well reviewed did not put a single one into ChatGPT's answer.
How many Columbus dentists show up in AI answers?
Almost none. ChatGPT named 0 of 58 Columbus practices. Perplexity named 3 of the 21 it could return a result for. Our Charlotte scan landed in the same place, with 2 of 154 named. A Columbus practice missing from AI answers is not an outlier right now. It is the default.
Here is what each engine named, and the base we measured it against. We tested ChatGPT and Perplexity. We did not test Gemini in this scan, so we do not publish a Gemini number.
| AI engine | Columbus practices it named | Base |
|---|---|---|
| ChatGPT | 0 | all 58 practices |
| Perplexity | 3 | the 21 it returned a readable result for |
| Gemini | Not measured in this scan | n/a |
| Either engine | at least 3 | only 21 could be read in both engines |
Source: Citevio scan of 58 Columbus, OH dental practices, June 19–23, 2026. Our Perplexity check returned an error for 37 of the 58, so its count is figured on the 21 readable results. ChatGPT was readable for all 58, which is why the zero is the number we stand behind.
A zero looks dramatic, so it is worth putting a national frame around it. SOCi's 2026 Local Visibility Index, which looked at nearly 350,000 business locations across 2,751 brands, put ChatGPT's local visibility rate at just 1.2% nationally. At that rate, in a group of 58 practices you would expect roughly one to show up. We found none. So Columbus is not broken in some special way. It is sitting at the low end of a range that is already close to the floor everywhere. The same study put the difficulty plainly: 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, January 2026. searchengineland.com
How we ran this scan
We took 58 Columbus dental practices and asked ChatGPT and Perplexity the questions a patient would actually type, then recorded whether each engine named the practice, a competitor, or no one. The scan ran from June 19 to 23, 2026. We publish the date because AI answers change, and a June result is a snapshot rather than a permanent verdict.
Most practices, 37 of the 58, were tested against these four patient-style queries:
- "best dentist in Columbus"
- "best cosmetic dentist in Columbus"
- "best Invisalign dentist in Columbus"
- "best dental implant dentist in Columbus"
The other 21 practices, scanned four days earlier, went through the same four questions with one defect we want to be upfront about: a bug in our scan template pasted a category label into the location slot, so those queries read "best dentist in Columbus Oh Cosmetic Dental Clinics" instead of "best dentist in Columbus." We are reporting it because it changes how much weight those 21 deserve. The headline does not rest on them. ChatGPT named none of the 37 practices tested with clean queries either, so the result holds on the clean batch by itself.
A practice counted as visible if an engine named it for at least one of its queries. We started from 72 scan files, removed duplicates, then removed five businesses that were not dental practices at all but shared a local listing (a stucco supplier, a woodworking store, a med spa and two plastic surgery clinics), which left 58 genuine practices.
One more limit worth stating. AI answers drift week to week: a 2025 Profound analysis of roughly 80,000 prompts per engine found that 40–60% of the sources an engine cites for the same question change within a single month. Treat these numbers as June's weather. The full method sits in how we measure AI visibility.
Volatility figure: Profound domain-drift analysis, July 2025 (~80,000 prompts per engine). tryprofound.com
What do the invisible Columbus practices have in common?
Not what most owners assume. They were not blocking AI, and they were not missing from the web. 54 of the 58 openly let AI crawlers in, and every Columbus site we could read was already in Bing's index. What almost all of them lacked was complete structured data and the kind of mentions elsewhere on the web that AI engines lean on.
Three signals came up again and again:
| Signal | Columbus practices | What it means |
|---|---|---|
| AI crawlers allowed in robots.txt | 54 of 58 (93%) | Access is not the problem; only 2 practices actually blocked AI bots |
| Complete structured data (schema) | 3 of 58 (5%) | Structured data helps engines read a page; 55 of 58 had it missing or incomplete |
| Present in Bing's index | 21 of the 21 we could read | ChatGPT pulls from Bing's results, and none of these sites were thin |
Source: Citevio scan, June 2026. Our Bing check hit a rate limit on 37 of the 58 sites, so the index figure covers the 21 we could read.
The robots number is the one that lands hardest. Of the 54 practices that openly welcomed AI crawlers, all 54 were still absent from ChatGPT. Letting the bots in is table stakes, not a strategy. Structured data, meaning the behind-the-scenes labels that tell an engine "this is a dentist, here are the hours and services," is a reading aid rather than a trigger, and it was missing or incomplete on 95% of the sites we checked.
So what does move the needle? Ahrefs studied 75,000 brands in 2025 and found that mentions of a brand across the web correlated with AI visibility at 0.664, while links pointing to the site came in at 0.218. Being talked about elsewhere tracked more than three times more closely with showing up than being linked to did. That is the layer nearly every practice in this scan was missing.
Correlation figures: Ahrefs study of 75,000 brands, May 2025. ahrefs.com
Does being in Bing's index get you into ChatGPT?
No. Every one of the 21 Columbus sites we could read was indexed in Bing, from about 50 pages to more than 145,000. ChatGPT named none of them. Being in the index is a requirement for showing up in ChatGPT. It is not a reason to be chosen.
This is where Columbus tells a different story than our Charlotte scan. In Charlotte, a thin Bing footprint was part of the picture: 22 of 154 practices looked genuinely under-indexed. In Columbus we found nothing thin at all. Thirteen of the 21 carried between 50 and 100 indexed pages, and eight were indexed in the tens of thousands of pages or more. On paper, every one of them was findable.
One caveat we owe you here. All 21 of these sites come from the earlier batch, the same one with the mislabelled location in its queries, because our Bing check hit a rate limit on the rest. So read this section as indicative rather than settled. What does not change across either batch is the outcome: no Columbus practice was named by ChatGPT at all.
They still lost. And the mechanism is not mysterious. As Damian Rollison described in Search Engine Land in 2025, 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." The bottleneck is not whether Bing has your pages. It is whether your pages come back in that top 20-30 for a patient's question, and whether the engine finds enough corroboration elsewhere to put your name in a recommendation.
ChatGPT's reliance on Bing per Damian Rollison, Search Engine Land, May 2025. searchengineland.com
Does a five-star rating help?
Not on its own. The Columbus practices ChatGPT skipped averaged 4.74 stars, with a median of about 205 reviews. 93% of them already sat at 4.3 stars or higher, the level AI engines tend to favor. Fifteen of them held a perfect 5.0. Every single one was still absent from ChatGPT's answers.
This is the finding owners find hardest to accept, because the reputation took years to build. The numbers below cover the practices ChatGPT did not name, which in this scan means all 58 of them.
Four invisible practices carried more than 1,000 reviews each. Eight sat at 4.8 stars or better with 700 or more reviews. If reviews alone decided AI recommendations, those would be the names every engine returns. They are on none of them. The reason is that ChatGPT already assumes a good dentist has good reviews, so once you clear roughly the 4.3-star line, more stars barely move anything. Google reviews still matter for patients and for the map pack. They are just not the lever for AI visibility once you are over the floor.
Three Columbus practices that should be everywhere, and aren't
To show what "invisible" looks like, here are three real Columbus practices from the scan, kept anonymous. All three were tested with our clean query set, all three openly allow AI crawlers, and none was named by ChatGPT for any of the four questions. The gap between their reputation and their AI visibility is this whole report in miniature.
The three practices that did surface, all of them in Perplexity, were not bigger names than these three. All three carried fewer than 800 reviews, and one of them surfaced with just over 100. That is the pattern worth sitting with: in Columbus right now, being findable beats being famous.
How does Columbus compare with Charlotte?
Both cities sit at the floor. ChatGPT named 2 of 154 Charlotte practices, or 1.3%, and 0 of 58 in Columbus. At this sample size a zero and a 1.3% are the same finding, so we do not claim one city is harder than the other. The review profiles are close too, which points at the same bottleneck in both: structured data and mentions, not reputation.
| Measure | Charlotte | Columbus |
|---|---|---|
| Practices scanned | 154 | 58 |
| Named by ChatGPT | 2 (1.3%) | 0 |
| Average rating of the invisible | 4.82★ | 4.74★ |
| Complete structured data | 3 of 154 (2%) | 3 of 58 (5%) |
Source: Citevio scans, June 2026. Charlotte scanned June 6–23, Columbus June 19–23.
It would be easy to sell you a story about Columbus being the tougher market. The math does not support it. A city running at Charlotte's 1.3% rate would still return zero across 58 practices about half the time, purely by chance. So the difference between these two columns is sample size, not difficulty. What both scans do show, consistently, is that neither market has a dental practice that AI engines reliably recommend. The full breakdown is in the Charlotte dentists report.
What can Columbus dentists do about it?
Start by seeing what a patient sees, then fix the plumbing before the polish. Check your own practice in ChatGPT and Perplexity, confirm your site is indexed in Bing, add the structured data that helps engines read you, and get your practice mentioned on sites other than your own. None of it involves paying an engine for placement.
Two next steps, in order:
- See where you stand today. Ask ChatGPT "best dentist in Columbus" and "best cosmetic dentist in Columbus" yourself, or read how to check what ChatGPT says about your practice for the step-by-step version.
- Fix what is keeping you out. Indexing, structured data, mentions and consistency, in that order. The full sequence is in how to get your dental practice recommended by ChatGPT.
We run this same scan in other cities. If you work outside Columbus, the Charlotte dentists report uses the identical method on a different market. Or check your own practice in about two minutes below.