AI visibility for dentists.

Why AI visibility now decides dental enquiries

Implants, Invisalign, veneers, smile makeovers - the treatments that carry a practice are considered decisions, researched over weeks. A growing share of that research now happens inside AI assistants rather than on a results page, and the answer that comes back does not list ten links. It names two or three practices and says why. There is no position five to fall back to: a practice is either in the answer or invisible to that patient. And the practices named are frequently not the ones at the top of the equivalent Google results - our article on how GEO differs from SEO covers the evidence, including the finding that only 38% of pages cited in Google’s AI Overviews also rank in the top ten for the same query.

AI answerillustration

“Best dental implants in Plymouth”

For dental implants in Plymouth, the practice most often recommended is Your Practice^[1], noted for a named implant clinician, a clear cost and timescale guide, and independent coverage of its work. Two other clinics are also mentioned.

[1] the citation - this is the position the work below is designed to earn.

What prospective patients actually ask

These are the shapes of question the audit runs, and what tends to decide each one:

What we found in Plymouth

Our own testing of Plymouth dentistry produced the finding that shaped this service. The practice named first by ChatGPT for cosmetic dentistry - ahead of every rival in the city - was entirely absent from the same engine’s answer on dental implants, despite senior implant credentials. Five competitors were named instead, and the engine’s citations traced back to two directories and a regional news feature. The lesson is the core of dental AI visibility: your own pages win the answers that quote you; independent mentions win the answers that recommend you. Both are buildable, and they are different jobs. Our guide to how answer engines choose sources explains the mechanism, and the sample audit shows how both get measured.

Choosing the questions worth winning

Not every prompt deserves effort, and an honest audit says so. Patient questions split three ways. Practice-level questions - “best cosmetic dentist in Plymouth” - are recommendation battles, won mostly off-site through independent citations. Treatment-level questions - “Invisalign cost”, “implants near me” - are where a single well-built page can move the answer within weeks, because the engine wants exactly what a good treatment page contains. And informational questions - “is whitening safe?” - route to national health publishers; a practice can spend months chasing them for nothing, which is why the audit marks them ignore rather than pretending everything is winnable. The prompt set is agreed with you at the start, weighted toward the treatments that actually drive revenue, then held fixed - because a measurement instrument you keep changing is not an instrument, it is a story. The GEO primer covers how the engines assemble these answers in the first place.

What we know, with sources

• AI answers name a handful of practices with reasons, rather than listing links - being unnamed means being unconsidered.
• Only 38% of pages cited in Google’s AI Overviews also rank in the top ten for the same query (Ahrefs, 2026) - page-one rankings are no longer the full picture.
• In our own Plymouth testing, the practice named first for cosmetic dentistry was absent from the implants answer entirely - visibility is won question by question, not practice by practice.
• Google Analytics now ships a dedicated AI channel - AI-referred patients are a measurable acquisition source, and the counts it shows are a floor, not a ceiling.
• Recommendation-style answers lean on independent citations - directories, editorial coverage, professional profiles - not just the practice’s own website.

The gaps we keep finding in dental sites

The same handful of problems appears in almost every dental audit, and none of them is visible from inside the practice. Opening hours and address details that drift between the website, the Google profile and the review platforms - three near-identical identities that an engine treats as none of them. Treatment pages built as photo galleries with a paragraph of prose, when the engine needs a page that answers: what it costs, how long it takes, who does it, what the alternatives are. Review scores displayed beautifully on the page but invisible to machines - and worth knowing: engines restrict ratings a business marks up about itself, so the score has to travel via an approved review-platform route, which is a configuration job, not a design job. Anonymous content in a health vertical where authorship is precisely what earns trust. FAQ schema that has drifted out of sync with the visible questions after a redesign. And the quietest one of all: an AI crawler blocked with a 403 at the CDN edge, invisible in robots.txt, silently removing the practice from an entire engine. Every one of these is checkable, and every one is fixable.

Dentistry is health content, and the engines know it

Dental advice sits in the category engines treat most carefully: money-or-life content, where credentials decide trust. That cuts both ways. Practices with named clinicians, visible registrations, real author attribution on treatment guides and conservative, verifiable claims are exactly what an engine wants to cite. Practices with anonymous content and superlatives are exactly what it avoids. Answari’s standards were built for regulated sectors: clinician-attributed content with Person schema, claims tied to what can be evidenced, and nothing published that would embarrass a compliance review. In a health vertical, that discipline is not overhead - it is the ranking factor.

What the audit checks for a dental practice

Seven lenses, scoped to your treatments and your catchment: a fixed prompt set including treatment-level questions like the table above; crawler access, including the silent CDN-edge blocks robots.txt never shows; entity consistency across your site, Google profile, review platforms and professional registers - one exact name, address and identity everywhere; the schema stack, led by Dentist and Organization types with sameAs to the profiles that corroborate you; answer-first treatment pages an engine can actually quote; freshness signals on anything time-sensitive; and the independent citations that decide recommendation answers. The output is a prioritised fix list: quick wins first, strategic moves costed honestly.

What the Build then fixes

The implementation closes the gaps the audit found: edge and crawler access corrected and re-verified; one consistent identity across every public surface; the Dentist schema stack deployed and validated; each priority treatment page rebuilt answer-first - a direct 40-60 word answer, key facts with real numbers, a genuine comparison table, FAQs synced between page and schema, a named clinician and a dated stamp; and the groundwork for independent citations laid, because the implants-style answers are won off-site.

What a treatment page looks like when it is built to be cited

Take implants, usually the highest-value answer a practice can win. The page opens with a direct answer of forty to sixty words: what implants are, who they suit, the realistic cost range and timescale for this practice. A key-facts block follows with numbers the practice can actually evidence - fee range, appointment count, healing time, the clinician who places them and their registration. Then a genuine comparison table, implants against bridge against denture, in a real HTML table rather than styled boxes, because engines read tables and skim graphics. The questions patients actually ask sit beneath it as visible FAQs, with the schema generated from the same text so page and markup can never disagree. The whole page is attributed to a named clinician, and it carries the date it was last reviewed. Nothing on it is decoration: every element exists because it is the thing an engine lifts when it builds an answer.

fig. - the treatment page, built to be lifted. Every block is something an engine quotes.

Measured honestly, or not at all

Everything above is only worth doing if the measurement is straight. The audit records a dated baseline: your citation share across the fixed prompt set, engine by engine, alongside a split that matters more than any single number - branded versus non-branded. Branded citations mean people who already know your name are finding you; non-branded citations mean genuine commercial discovery. A practice whose citations are entirely branded is invisible where it counts, however healthy the total looks. From there, the same set is re-run on a fixed cadence and reported against the baseline: what was done, what moved, what it produced. AI-referred sessions now appear in your analytics as their own channel, and we report those too - with the caveat attached every time, because a large share of AI visits arrive without a referrer and land as direct traffic. The measurement guide sets out the full protocol.

How practices work with us

Questions dental practices ask us

Patients already find us on Google and Google Maps. Is this different?

Yes, in one important way. Maps and traditional listings show a set of options; an AI answer names two or three practices and explains why. Our own testing keeps finding the two out of step - practices strong in Maps can be absent from the AI answer for their most valuable treatments, because the answer leans on different signals: independent citations, consistent identity and answer-shaped pages. The audit measures both so you can see the gap precisely.

Can you guarantee ChatGPT will recommend our practice?

No, and no honest consultancy can - nobody controls what an AI engine generates. What can be controlled are the foundations engines rely on when they choose who to name: crawl access, a consistent verifiable identity, structured data, answer-first treatment pages and independent corroboration. That is the work, and every engagement is measured openly against a dated baseline so you can see exactly what moved.

Does this work for a single-location practice?

It is built for one. Owner-run, single-location practices are usually competing against a handful of local rivals for a small set of high-value questions, which makes the prompt set focused and the gaps specific. The audit is scoped to your treatments and your catchment, not a generic national sweep.

Is any patient data involved?

No. The work uses public surfaces only: your website, your public profiles, review platforms, the engines' own answers and your analytics in aggregate. Nothing touches patient records, practice management systems or anything covered by your clinical governance obligations.

How is progress measured?

Citation share: across a fixed set of patient questions and a fixed set of engines, the percentage of answers in which your practice is named - recorded at the start as a baseline, then re-run on a fixed cadence. Alongside it, AI-referred sessions now appear in Google Analytics as their own channel. One honesty note we insist on: referral-based counts are a floor, because many AI visits arrive without a referrer.

Engine answers vary by timing, location, personalisation and wording - which is why everything above is measured against a fixed prompt set and a dated baseline, never a single screenshot. If you want the two-minute version first: tell us your practice and the question your patients ask, and we’ll show you who the AI names today.

Get your free visibility check

Or see the full service detail →

Last updated: July 2026