Health practices are among the weaker sectors we measure. The 70 dental practice websites in our corpus average 52 out of 100 on the signals AI assistants read, with 37% under 50 — and physicians fare worse still, averaging 48 with 46% under 50, near the bottom of the 2,528-site corpus. When a new patient asks an assistant for a dentist nearby who takes their insurance, most practice websites give it very little to answer with.
The dentist cut splits cleanly into haves and have-nots: 56% had schema.org business data, 66% had SPF on the domain — both above corpus average — yet 39% had no meta description at all, the sentence an assistant would quote to describe the practice. For context inside medicine: doctors averaged 48 with 41% schema adoption, and pharmacies were the corpus's worst cut at 46 average with 60% of sites under 50.
The questions are concrete — where, when, what number — and the crawler answers them from your raw HTML only. A practice site built as a JavaScript application delivers an empty shell to GPTBot and ClaudeBot; a populated LocalBusiness block with name, full postal address, telephone and opening hours makes those facts quotable in one place. The meta description matters more for practices than most businesses, because it is where the assistant finds what kind of dentistry you do.
View your homepage with JavaScript disabled and confirm the address, phone number and hours survive. Search your page source for application/ld+json and check the fields are filled, not empty. Read robots.txt for AI crawlers blocked by name. Or run the free scan — it measures all of it in about a minute, makes at most thirteen read-only requests, and runs only because you asked it to.
Slightly below the middle: an average of 52 out of 100 across 70 scanned dental sites, against a 53 corpus average — but with above-average schema adoption at 56%. Doctors and pharmacies score lower, at 48 and 46.
Because it is the one sentence in your HTML that says what kind of practice you are, and 39% of the dental sites we scanned had none. An assistant describing you has to quote something; with no description and no populated schema, it has nothing.
No. The scan reads only what your site publicly serves — the homepage's raw HTML, robots.txt and public DNS records. It does not log in, submit forms, or go beyond the homepage's own links.
One populated LocalBusiness JSON-LD block — name, full address, telephone, opening hours, url — in the served HTML, plus a real meta description naming your services and neighbourhood. Those are the exact fields the scan reads one by one.
Where does your site stand? The scan is free, takes about 60 seconds, and needs no account. It fetches your homepage exactly the way GPTBot, ClaudeBot and PerplexityBot do and tells you what they receive.