We checked the AI visibility of 16 dental practices. Three were invisible to AI.
Customers now ask ChatGPT and Google's AI "who's a good dentist near me?" So we took 16 real dental practices across Austin, TX and Columbus, OH and scored how well an AI assistant can actually understand and recommend each one from its website.
We ran every homepage through a tool that reads a site the way an AI assistant does and scores it from 0 to 100 on the things a model needs to recommend a business: what it does, where it serves, its services, structured data, contact details, pricing signals, hours, answers to common questions, and trust signals. Dentistry is a high-value, marketing-savvy niche, so we expected strong results. Most were solid. But the most important finding wasn't a low score.
Nearly one in five was simply invisible
Three of the sixteen practices couldn't be read at all. Not a low score, no score: the tools that feed AI answers loaded the page and found nothing usable, almost always because the site is built entirely in JavaScript or sits behind bot-protection that blocks the crawl. When an AI assistant, or the crawler behind an AI answer, hits a wall like that, the business isn't scoring a C. It's not in the conversation. A patient asking an AI for a dentist will simply never hear its name.
This is the quiet killer. These practices likely look fine to a human in a browser, so nobody notices. But a growing slice of "who should I see?" now happens inside an AI that never gets past the front door. Invisible beats a bad score, and not in a good way.
The readable sites: findable, not always the one AI picks
Of the thirteen that could be read, the average was 73, and ten of thirteen had structured data in place. These are practices that have done the basics. Here's how the full set landed:
Three practices reached an A, and the top site scored a 96, so it's clearly achievable. But most cluster at B and C: good enough to appear, not always good enough to be chosen.
Where the readable sites lost points
The pattern was consistent. AI could tell what these practices do and roughly where they are. What it repeatedly couldn't pin down was the deciding detail: which insurance they take, whether they see new or emergency patients, rough pricing, exact hours, and plain answers to the questions people actually ask before booking. That is exactly the information an assistant needs to recommend one dentist over another, and when it isn't in readable text, the model hedges and names a competitor it understood better.
The three lowest readable scores had one thing in common: no structured data. Every site missing JSON-LD landed in the bottom of the pack, and the single D-grade site had neither structured data nor clear specifics. The schema that hands AI your facts directly is the difference between a model guessing about you and quoting you.
Check your own practice's AI visibility — free →The takeaway
Even in a niche with money and marketing behind it, nearly one in five dental sites was invisible to AI, and the readable average was a 73. Being "findable" on Google was the old bar. As patients shift to asking AI directly, the new bar is being readable at all, and then being the practice the AI is confident enough to name. The first is a technical fix; the second is a weekend of clearer content and structured data, not a rebuild.
See where your practice scores →More from this series: plumbers, HVAC companies, personal injury firms, and med spas. Also: why ChatGPT doesn't recommend your business and how to show up when customers ask AI for a business like yours.
Method: 16 dental practice homepages across Austin, TX and Columbus, OH, scanned in September 2026 with BotDesk's AI Visibility Checker. A spot-check, not a census; individual practices are not named.