The AI Visibility Gap in UK Dental Practices: What 200 Surgeries Are Missing in New Patient Revenue
Joshua Etheridge
Only 14% of UK dental practices appear in AI search recommendations. When AIO audited 200 practices across ChatGPT and Google Gemini in mid-2026, fewer than one in seven was named when a prospective patient asked an AI assistant to recommend a dentist. The other 86% were absent at the moment of highest purchase intent.
The scenario is now ordinary. A prospective patient opens ChatGPT and types "best NHS dentist near me accepting new patients", or asks Gemini to recommend a private dental practice in their city. A small number of practices get named. The vast majority do not, and the patient revenue sitting in that gap is, on a per-practice basis, worth paying close attention to.
This is not a story about a distant future technology shift. UK patients are already turning to AI for health decisions. A March 2026 study from King's College London and its Policy Institute, surveying more than 2,000 UK adults, found that one in seven (15%) had used an AI chatbot for health advice instead of contacting a GP or other NHS service (King's College London). Dental care, with its combination of high out-of-pocket costs, strong geographic intent, and real anxiety around provider choice, is exactly the kind of category where an AI recommendation is likely to carry weight. When a patient asks an assistant for a shortlist, a practice that does not appear is invisible at the moment of highest purchase intent.
Methodology
AIO researchers pulled practice data across 200 UK dental surgeries spanning fourteen cities: London, Manchester, Birmingham, Leeds, Bristol, and Edinburgh, plus eight further cities. Each city contributed between roughly 12 and 16 practices to the sample. Practices were drawn from Google Places data, covering both NHS-accepting and private practices, from single-chair independents through to multi-site groups.
For each practice, across late May and June 2026, we ran a structured prompt battery against the consumer ChatGPT and Google Gemini assistants as they shipped at the time of testing, including their live browsing and retrieval layers. We describe them as products tested on a date rather than by underlying model version on purpose, because for local recommendations it is the browsing and retrieval layer, not a base model weight, that decides which practices get named. AI citations are also volatile, so we ran every prompt five times per assistant across the test window and judged each practice on how it performed across those runs rather than on a single result. Prompts covered three intent types: generic local discovery ("recommend a dentist in [city]"), treatment-specific queries ("where can I get Invisalign in [city]"), and trust-signal queries ("which dentists in [city] have the best reviews"). We recorded whether each practice was cited by name, cited as a category reference without naming, or absent entirely. We also assessed each practice's underlying digital footprint, reviewing Google Business Profile completeness, review volume and velocity, structured data on their website, presence on third-party directories including Trustpilot, Whatclinic, and Yell, and the quality of their editorial web presence.
One note on sourcing. The citation findings above are AIO's own primary research, measured directly in the prompt battery. The revenue figures later in this article are a different kind of number. They are AIO estimates, built from publicly published NHS dental charge bands and from commonly cited industry ranges for private fees, enquiry volumes, and conversion rates. We have not attributed them to a single named third-party study, because we would rather label an estimate honestly than dress it up as an external finding. Throughout, treat the audit data as measured and the revenue model as illustrative.
Key Findings
AI citation rates are very low across the board
Across our 200-practice sample, only 14% appeared in any AI-generated recommendation during the prompt battery. Of those 28 practices, just nine were cited consistently across both ChatGPT and Gemini, appearing in the majority of runs on multiple related queries rather than on a single lucky match. The remaining 19 appeared on one platform or one query type only, which matters because single-platform presence is fragile. Model updates, retrieval changes, and competitor activity can dislodge it quickly.
Put differently, 86% of the practices we audited were entirely absent from AI search at the point a patient was actively seeking a recommendation. That is not a marginal visibility problem. It is a near-total absence from a growing acquisition channel.
The practices that do appear share a clear profile
The nine consistently cited practices were not necessarily the biggest or the oldest. What they had in common was a combination of four measurable factors. First, high review volume: all nine had more than 120 Google reviews, with an average of 214. Second, strong review recency: each had received at least one new review within the previous 14 days at time of audit. Third, structured data: all nine had either LocalBusiness or MedicalOrganization schema correctly implemented on their website. Fourth, third-party citation depth: all nine appeared on at least four dental-specific or healthcare-specific directories with consistent NAP (name, address, phone) data.
The practices that appeared on only one platform or one query type typically had one or two of these factors but not all four. The 86% that were absent had, on average, 41 Google reviews, no schema markup, and appeared on fewer than two external directories beyond Google itself.
| Signal we measured | Consistently cited (9 of 200) | Occasionally cited (19) | Absent (172) |
|---|---|---|---|
| Google review volume | 120+ (avg 214) | Moderate | ~41 on average |
| New review in last 14 days | All nine | Some | Rarely |
| LocalBusiness / MedicalOrganization schema | All nine | Some | None |
| Health/dental directories with consistent NAP | 4 or more | 1 to 3 | Fewer than 2 |
Treatment-specific queries produce the worst results for most practices
Generic local discovery queries ("dentist in Manchester") returned at least some named practices in 78% of our test runs. Treatment-specific queries ("Invisalign provider in Bristol", "dental implants Leeds") returned named practices in only 31% of runs. For high-value elective treatments, which carry the highest revenue per patient, AI assistants are most likely to give generic category answers or redirect to aggregator sites rather than naming a specific practice.
This is a structural gap. Practices with strong treatment-page content, clear pricing information, and specific structured data for individual services dramatically outperformed on these queries. Of the nine consistently cited practices, seven had dedicated landing pages for their top three treatments with schema-marked pricing ranges. Among the absent 86%, that figure was under 4%.
The Revenue Model: What the Gap Costs Per Practice
To make this concrete, here is an illustrative revenue model at the individual practice level. It is a worked example rather than a forecast, and its single biggest lever, the share of new enquiries a practice could realistically win back through better AI visibility, is an assumption we flag explicitly below rather than a guaranteed return.
As an AIO working estimate, the average UK dental practice treats on the order of 1,200 active patients and generates somewhere between 35 and 55 new patient enquiries per month across all channels combined, of which we assume roughly 60% convert to a first appointment. For patient values we use blended AIO estimates: a new NHS patient worth on the order of £280 across an initial course of treatment, taken as a blend across the published NHS charge bands, a new private patient around £620 for an initial course, and elective private treatments around £2,400 per case. Every figure in this paragraph is an estimate, not a third-party-audited number.
AIO estimates that AI assistants currently influence in the region of 8 to 12% of new patient discovery journeys in UK dentistry. This is our own planning estimate, not a measured market share. It is informed by findings like the King's College London study above on how many adults now use AI for health questions, together with what we see in audit data, and it should be read as an assumption rather than a precise figure. We expect the share to keep rising as AI answers become more embedded in Google Search and standalone assistant use grows, though the exact pace is uncertain.
Here is the assumption to scrutinise, because it drives everything that follows. We model a scenario in which moving a practice from absent to consistently cited lifts its total new enquiry volume by around 10%. Treat that as a scenario input, not a derivation. It is deliberately not the same number as the 8 to 12% AI-discovery share above. Being cited captures only a fraction of AI-driven discovery, not all of it, so we are not converting that share into enquiries one for one. The 10% is plausible rather than proven, and the honest range around it is wide. On that scenario, a mixed NHS/private practice generating 40 new enquiries per month gains four additional enquiries monthly. At a 60% conversion rate, that is 2.4 new patients per month, or roughly 29 per year. Blending NHS and private first-visit values at a 50/50 split gives an average first-visit value of about £450, so those 29 patients are worth approximately £13,000 in first-visit revenue in year one.
That figure does not include patient lifetime value, and the retention numbers here are again AIO estimates. Taking typical return rates of roughly 1.8 NHS visits and 2.3 private visits per year, and valuing a retained private patient with occasional elective treatment at £1,400 to £2,800 per year, the lifetime value uplift from those 29 new patients works out, on these assumptions, at somewhere between £38,000 and £62,000 over a three-year horizon for a typical mixed practice.
What AI Visibility Patterns Tell Us About How Models Rank Local Health Providers
The findings from this audit align with what AIO has observed across other high-trust local service categories including legal, financial, and optical. We want to be careful about mechanism here, because it is easy to overclaim. What we can say from the data is that the practices named consistently are not simply the oldest or the largest. In today's consumer AI assistants, local recommendations are heavily grounded in real-time retrieval, the layer that pulls in live web and profile data, rather than recalled purely from training weights. The practical implication is that the quality and legibility of what a model can retrieve about you matters at least as much as whether you exist.
For dental practices, this means three things are true simultaneously. Google Business Profile optimisation still matters because it feeds Gemini's local retrieval layer directly. Editorial web presence, meaning blog content, press mentions, local news coverage, and professional association listings, shapes how language models weight a practice during generation. And structured data on the practice website functions as a legibility signal, telling the model exactly what type of entity it is, where it operates, what it treats, and how it has been rated.
Review velocity is more important than review volume alone
One finding worth isolating: in our sample, the difference between practices cited consistently and those cited occasionally was not simply total review count. It tracked with the rate of new reviews arriving. In our audit, practices with around 90 reviews and four in the last month tended to outperform practices with 140 reviews and none in the last six months. We are describing a correlation we measured, not a claim about model internals, but it has a plausible mechanism: review recency is a known signal in the local retrieval layer these assistants draw on, and a profile that looks dormant reads as a weaker, less current recommendation regardless of historical star rating.
This is why review velocity campaigns are not just reputation management tools. They are an AI visibility input with direct citation consequences.
What Dental Practices Should Do
Based on the audit findings, the gap between cited and absent practices closes through a specific combination of actions rather than any single fix.
The first priority is structured data implementation. Adding correct LocalBusiness and MedicalOrganization schema to the practice website, with accurate service listings and geo-coordinates, is the highest-leverage technical change most practices can make. Fewer than 6% of the 200 practices audited had any schema markup correctly implemented.
The second priority is citation depth. Consistent NAP listings across Whatclinic, Healthgrades UK, Yell, Trustpilot, the NHS Choices directory, and relevant local business directories build the entity footprint that AI models use to confirm a business is real, established, and trustworthy.
The third priority is review velocity. Not a one-time push, but a sustained programme that keeps new reviews appearing on a regular cadence. A practice aiming for AI visibility should target a minimum of two to three new Google reviews per week.
The fourth priority is treatment-specific content. Dedicated pages for high-value treatments, written with clear clinical information, transparent pricing ranges where possible, and structured data marking the specific service type, dramatically improve performance on the treatment-specific AI queries that carry the highest patient value.
Closing the Gap Is a Timing Decision
The 86% of UK dental practices currently absent from AI search recommendations are not in a stable position. The channel is growing. Patient behaviour is shifting. And the practices in the cited 14% are compounding their advantage with every week that passes, accumulating more reviews, more citations, and more embedded presence in the models that will answer the next generation of patient queries.
The worked example points, on the assumptions set out above, to roughly £13,000 in first-year new patient revenue for a typical mixed practice and a three-year value in the tens of thousands. Treat those as an illustration of the stakes rather than a promise. What is not an assumption is the audit finding behind them: 86% of the practices we tested were absent from AI recommendations entirely, at the exact moment patients were asking for a name.
Frequently asked questions
How many UK dental practices appear in AI search?
In AIO's mid-2026 audit of 200 UK practices, only 14% appeared in any AI-generated recommendation across ChatGPT and Google Gemini, and just nine practices (4.5%) were cited consistently across both assistants.
What makes a dental practice get cited by AI assistants?
In our sample, consistently cited practices shared four measurable traits: more than 120 Google reviews, at least one new review in the prior 14 days, correct LocalBusiness or MedicalOrganization schema, and consistent listings across four or more health or dental directories.
Why do treatment-specific queries perform worse?
Generic prompts such as "dentist in [city]" returned named practices in 78% of runs, but treatment queries such as "Invisalign provider in [city]" named a specific practice in only 31% of runs. Practices with dedicated, schema-marked treatment pages performed far better.
Is the revenue impact guaranteed?
No. The citation findings are measured directly in the audit, but the revenue figures are an illustrative AIO estimate built on stated assumptions, the largest being a roughly 10% uplift in total enquiries from moving to consistent citation. Treat them as a scenario, not a promise.
How is AI visibility different from traditional SEO?
Traditional SEO competes for a place in a list of links. AI visibility decides whether your practice is named at all inside a generated answer, and it leans on review recency, structured data, and directory and entity signals as much as on search rankings.
Find out exactly where your practice stands. AIO's free audit analyses your current AI citation rate across ChatGPT and Gemini, identifies the specific gaps in your digital footprint, and gives you a prioritised action plan with estimated revenue impact. It takes two minutes and costs nothing. Run your free AIO audit and see what AI search is currently saying about your practice.