We Checked 100 NYC Dental Websites for AI Readiness. Almost Half Scored Below 60.

What we found

The good news first: almost no practice is actively shutting AI out. Only 3% of sites block major AI crawlers in their robots.txt file, and 95% use HTTPS. The problem isn’t that AI is locked out — it’s that once AI gets in, most homepages don’t tell it much it can use.

The three most common gaps

1. No answers to the questions patients actually ask

52% of homepages had no FAQ or question-and-answer content at all. Another 42% had some question-style content but without the structured markup that lets AI read it as a clear set of questions and answers. Only 6% had both.

Why it matters: People ask AI questions — “Does this dentist take new patients?”, “Do they do Invisalign?”, “Is there emergency care on weekends?” AI answers most confidently from pages that already answer those questions directly.

How to fix it: Add a short FAQ section with 5–8 real questions your front desk hears every week, and mark it up with FAQPage schema. For most sites, this is an afternoon of work.

2. Nothing that tells AI why this practice is trustworthy

Only 9% of homepages stated two or more concrete credibility signals in their text — things like years in practice, board certification, awards, or media mentions. Most had basic contact details or an About link, but nothing specific that sets the practice apart.

Why it matters: When AI names a handful of practices, it leans toward ones it has clear reasons to trust. “Serving Park Slope since 2004” or “board-certified periodontist” is the kind of detail it can repeat.

How to fix it: Write your credentials into the page as plain text — not only inside images or logos.

3. No structured data telling AI what the business is

29% of sites had no structured data (Schema.org / JSON-LD) at all. Structured data is a small block of code that states, in a format machines read reliably, “this is a dental practice, at this address, with these hours and this phone number.”

Why it matters: Without it, AI has to infer those basics from the page’s design and wording — and it can get them wrong or skip the practice entirely.

How to fix it: Add Dentist (a LocalBusiness type) schema with your name, address, phone, and hours. Many website platforms and SEO plugins can generate it.

Borough by borough

The differences between boroughs are small and each group has only 19–20 practices, so this shouldn’t be read as a meaningful ranking. The bigger story is how similar the picture is everywhere: in every borough, the typical practice lands in the mid-50s.

What a practice can do this week

1. Add an FAQ section with the questions patients really ask, marked up with FAQPage schema.
2. Add Dentist schema with your name, address, phone, and hours.
3. Put your phone number and address in plain text, not only in an image or a map embed.
4. State your credentials in words — years in practice, certifications, specialties.
5. Check that your main content is in the page itself. Open your site, choose “View Page Source,” and search for a sentence from your homepage. If it isn’t there, AI crawlers that don’t run JavaScript probably can’t see it either.
6. Run a free check of your own site at gocited.pro/ai-visibility to see which of these signals you already have.

Methodology

Sample. On September 24, 2026, we used Google Places Text Search with the query “dentist in [Borough], New York, NY” for each of the five boroughs and took the first 20 results that listed a website, in Google’s order. We excluded national and regional chains (3), duplicate websites belonging to the same practice (3), and results without a website (1). That produced 100 independent practices. Three sites could not be reached by our checker (they blocked automated requests or timed out), leaving 97 in the results.

What we measured. Each homepage was scored from 0 to 100 on 13 structural signals using GoCited’s public AI-Readiness Checker: structured data (20 points), content depth (15), FAQ / answer-first content (15), credibility signals (10), llms.txt (8), AI crawler access in robots.txt (7), heading structure (6), image alt text (5), canonical tag (4), HTTPS (4), page title (3), meta description (2), and Open Graph tags (1).

How pages were read. The checker reads the HTML a site delivers without running JavaScript. That matches how the crawlers behind ChatGPT, Claude, and Perplexity work: a Vercel and MERJ analysis of more than 500 million GPTBot requests found no evidence of JavaScript execution, and the same held for ClaudeBot and PerplexityBot. Google’s Gemini is an exception — it renders JavaScript through Google’s own infrastructure — so these results best reflect what non-rendering AI crawlers see.

Frequently asked questions