Dental · Primary care · Optometry · Chiropractic · Specialty practices

To an AI engine, your practice might be three different businesses.

Healthgrades knows your dentists as individual clinicians. Google knows your practice as a place. Your website presents a brand that may match neither. Every other local business has one identity problem to solve. Healthcare practices have three overlapping ones — and an engine that can't reconcile them simply names someone else.

Check what AI says about your practice — free

No account, no card required.

The Core Problem

Three records, one business, no link between them

IDENTITY 1 — THE CLINICIAN

Provider directories index people

Healthgrades, Vitals, Zocdoc, WebMD Health, and 1-800-Dentist all build their records around the individual practitioner, tied to credentialing data. Reviews attach to the clinician. If a dentist has worked at three practices, that history can follow the person, not the practice — and it may still name a former employer.

IDENTITY 2 — THE PLACE

Google Business Profile indexes a location

A profile with the practice name, an address, hours, and the review history. Multi-location groups get one per site, which is correct — but each one needs its own consistent details, and the suite number is a surprisingly common point of disagreement.

IDENTITY 3 — THE BRAND

The website presents something else again

Often a marketing name ("Smile Bright Family Dental") that isn't the legal entity, sitting on a site whose provider pages never state which address that provider actually works at. The site is the one record you fully control, and it's usually the one doing the least to connect the other two.

An AI assistant asked to recommend a dentist has to decide whether these are one practice or three similar ones near each other. When it can't tell, it doesn't guess — it falls back to a directory listing, which is the safest answer available. Your practice is in that listing somewhere, in a list of thirty.

The fix is not to pick one identity. It's to state the connection explicitly: identical practice name, address, and phone across every record, and provider pages on your own site that say plainly which practice that clinician works at and at which address. Stated links beat inferred ones, because an engine can quote a stated link.

The Biggest Gap

The first question every patient asks is the one your site can't answer in text

Before location, before reviews, before anything: do they take my insurance? It's the filter that eliminates most of the list. And on the overwhelming majority of practice websites, the answer exists as a strip of carrier logos, a line saying "we accept most major plans," or a downloadable PDF.

None of those are facts a machine can read. A logo is an image. "Most major plans" is unfalsifiable and gets treated accordingly. A PDF is a wall. So when someone asks an assistant which dentists near them take a specific plan, the assistant goes to the payer's own provider directory or to Zocdoc — and the practice named is whichever one is listed there in a machine-readable way.

A plain text page listing each plan and network by name, stating in-network versus out-of-network for each, and explaining what happens if a plan isn't listed, is one of the highest-value pages a practice can publish. Very few have one.

One honest caveat: plan participation changes, and a stale insurance page is worse than no page at all — it produces confidently wrong answers about you. This only works with a last-reviewed date on the page and a real commitment to keeping it current.

The records VizAI checks for a practice

Healthgrades Zocdoc WebMD Health Vitals 1-800-Dentist US News Health

Alongside the seven general identity records — Google Business Profile, Yelp, Bing Places, Apple Maps, Foursquare, the BBB, and Facebook — checked for whether the practice name, suite-level address, and phone agree with each other and with your site.

Realistic Targets

AI won't call anyone the best dentist. Aim at the questions it will answer.

Health topics get handled cautiously by every major AI assistant. Ask for "the best dentist in [city]" and you'll frequently get a directory, a hedge, or a suggestion to check with your insurer — not a name. Chasing that query is chasing an answer the engine has decided not to give. The logistical and access questions are a different story: those get answered with specific practices all the time.

USUALLY DEFLECTED
  • "who is the best dentist in [city]"
  • "which orthodontist should I choose"
  • "is [procedure] right for me"
  • "which dentist is better, A or B"
ANSWERED WITH NAMES
  • "dentist in [city] accepting new patients"
  • "emergency dentist open Saturday near [town]"
  • "dentist near me that takes [plan name]"
  • "how much is a crown without insurance in [city]"
  • "pediatric dentist for an anxious child in [town]"
  • "dental office with early morning appointments [area]"

What that means for the site

Hours as data, not as a graphic

"Open Saturday" and "early appointments" are inside the query. Opening-hours markup is part of what VizAI scores, and in this vertical it directly answers half the question being asked.

The specific entity type

Schema.org has Dentist, Physician, MedicalClinic, MedicalBusiness, and Optician as distinct types. Using the right one scores higher than a bare Organization block, because only the subtype identifies what kind of local business this is. Provider pages should additionally carry Person markup.

Accepting-new-patients status, in writing

A remarkably common query, and a remarkably rare on-page fact. If a practice's panel is open, saying so plainly on a page — per provider, per location — is cheap and directly citable.

Cost ranges you're willing to publish

"How much does X cost without insurance in [city]" is asked constantly and answered almost entirely by national averages, because practices won't put numbers on a page. A published range with its assumptions stated is a genuine competitive opening.

What we don't write

No clinical content. Nothing that reads as treatment advice, diagnostic guidance, or a health claim. Content produced for a practice stays on access, logistics, cost, and process, and the practice reviews everything before it's published. That line is partly ethics and partly strategy — the logistical questions are the ones AI assistants will actually answer with a business name.

How the audit reports it

We ask Claude, ChatGPT, Gemini, and Perplexity the questions your patients ask, record which practices get named, and score the result on a 0–100 scale — 0–24 Invisible, 25–49 Emerging, 50–74 Present, 75–100 Dominant. Alongside the score you get every identity mismatch we found, named record by named record, and a prioritized list of what to fix first. Our team does that work; nothing arrives as a checklist for your front desk to absorb.

Questions from practices

Patients ask AI which dentists take their insurance. Can that be fixed on our website?

It can, and it's usually the single largest missed opportunity on a practice site. Insurance is the first filter almost every patient applies, and on most sites the answer exists as a logo strip, a line saying "most major plans accepted," or a PDF — none of which are readable as facts. A page naming each plan and network in text, stating in-network versus out-of-network for each, and explaining what happens for plans not listed, gives an assistant something it can use. Participation changes, so this only works with a last-reviewed date and a real commitment to keeping it current; a stale insurance page is worse than none.

Healthgrades lists our dentists individually but Google lists the practice. Which one should match?

All of them — this mismatch is the defining AI-visibility problem in healthcare. Provider directories like Healthgrades, Vitals, and Zocdoc index individual clinicians. Google Business Profile indexes the practice as a place. Your website usually presents a third thing, a brand name matching neither. That leaves an engine with three overlapping entities and no reliable way to tell they're one business. The fix isn't picking one: make the practice's name, address, and phone identical everywhere, and have every provider page on your own site state which practice that provider works at, at which address, so the individual and practice records are explicitly linked.

Will an AI assistant ever name a specific dentist as the best one?

Often it won't, and that's worth planning around rather than fighting. Health topics are handled conservatively, so a superlative question about a clinician frequently returns a directory, a hedge, or advice to consult a professional. That doesn't make the category unwinnable — it means the winnable questions are logistical: who's accepting new patients, who's open Saturday, who takes this plan, what a procedure typically costs without insurance, whether a practice sees children, what the new-patient process involves. Those get answered with specific businesses far more readily.

We just rebranded after acquiring another practice. What breaks?

Practically everything an engine uses to identify you — this is the highest-risk moment in a practice's visibility. After an acquisition there are typically two legacy names in circulation, one or two old numbers still forwarding, directory records under both names, review history attached to the old Google listing, and insurance directories updating on the payers' timeline rather than yours. Engines see several similar practices at overlapping addresses and get less confident about all of them. The work is unglamorous: reconcile name, address, and phone across every record, keep rather than abandon the listing holding the review history where you can, and give the old name an explicit place on the new site so the connection is stated rather than inferred.

Does VizAI write clinical or medical content for practices?

No. Content produced for a practice stays on access, logistics, cost, and process — hours, locations, insurance and payment, new-patient onboarding, what an appointment involves, price ranges where the practice is willing to publish them. We don't write treatment advice, diagnostic content, or clinical claims, and nothing about a practice is published without that practice reviewing it. The boundary is partly ethical and partly practical: the logistical questions are both safer and the ones an AI assistant will actually answer with a specific business name.

Different industry?

Each category has its own query grammar and its own directory layer. Three more we've written up:

Comparing this to the SEO work you're already buying? AEO versus traditional SEO, side by side. Agencies serving several practices can run this under their own brand — see the white-label program.

Find out which version of your practice AI knows

We'll check all three identities, ask the four major engines your patients' questions, and show you what came back. Free to start — no credit card.