Study: We Tested the AI Visibility of 50 Dental Practices in Paris — The Results

In March 2026, Eddie Miller Agency conducted a systematic audit of AI visibility for 50 dental practices spread across 12 Paris neighborhoods. The goal: to concretely measure how much (or how little) Parisian dentists are recommended by AI assistants when a patient asks a question.

The results are striking. The vast majority of dental practices in Paris are completely absent from AI recommendations. And the few that do appear share very specific characteristics.

Here are our complete findings.

Methodology

We selected 50 dental practices across 12 Paris arrondissements (1st, 5th, 8th, 9th, 10th, 11th, 12th, 14th, 15th, 16th, 17th, 20th), mixing independent practices, centers belonging to networks (Dentylis, Dentego, Dentelia, GranDental, Vertuo Sante), and practices of varying sizes (solo to multi-practitioner).

For each practice, we tested its visibility on three AI platforms: ChatGPT (chat.openai.com), Perplexity (perplexity.ai), and Google Search with AI Overview enabled.

We used 7 standard queries per arrondissement, phrased as a real patient would ask them: “Who is the best dentist in [neighborhood]?”, “I’m looking for a dentist for a dental implant in Paris [neighborhood], which one do you recommend?”, “Dentist emergency [neighborhood] reviews”, “Best orthodontist Paris [neighborhood]”, “Recommended dental practice near [local landmark]”, “[Practice name] dentist reviews Paris”, and “Compare dentists [neighborhood] implant price.”

Each query-platform combination was tested and documented. For each response, we noted whether the practice was mentioned, its position in the response (1st, 2nd, 3rd mention), what information the AI cited, and which sources were referenced.

Each practice’s GEO score was then calculated out of 100 points, according to our proprietary 15-criteria grid, divided into 5 categories: mention frequency (0-30 points), accuracy of cited information (0-20 points), position in recommendations (0-20 points), source diversity (0-15 points), and relative performance against competitors (0-15 points).

Overall Result: An Average Score of 18/100

The average GEO score across the 50 audited practices is 18 out of 100.

This figure means that the average dental practice in Paris is virtually invisible to AI assistants. When a patient asks an AI for a dentist recommendation, the vast majority of Parisian practitioners have no chance of being cited.

To put this result in perspective, international studies confirm the same trend. According to the SOCi 2026 Local Visibility Index report, which analyzed nearly 350,000 businesses, only 1.2% of local businesses are recommended by ChatGPT, compared to 35.9% appearing in Google’s Local Pack. AI visibility is 3 to 30 times harder to achieve than traditional local search rankings.

Score Distribution: Three Categories of Practices

The 50 practices fall into three very distinct groups.

The Invisible (score 0-15) — 62% of audited practices. These 31 practices don’t appear in any AI response, regardless of the platform or query tested. AI simply doesn’t know them, or doesn’t have enough reliable data to recommend them. Most share a similar profile: incomplete Google Business listing, few recent reviews, no website or a website without structured data, and a presence limited to one or two platforms.

The Ghosts (score 16-40) — 28% of audited practices. These 14 practices appear sporadically — cited once or twice on one platform, absent from the others. Often, it’s a single very specific query (their practice name + “reviews”) that triggers a mention, but generic queries (“best dentist [neighborhood]”) don’t surface them. The information cited by AI is sometimes incomplete or inaccurate.

The Visible (score 41-100) — 10% of audited practices. Only 5 practices out of 50 score above 40, with a maximum of 72/100 for the best-performing audited practice. These practices appear regularly in AI responses, often as the first or second mention, with accurate information and diversified sources.

What Sets the Visible 10% Apart from the Invisible 62%

The gap between visible and invisible practices is not linked to quality of care. It’s linked to very concrete digital signals. Here are the factors we identified, ranked by correlation with GEO score.

Factor 1: Volume and Freshness of Google Reviews

This is the factor most strongly correlated with GEO score in our study. The 5 visible practices have an average of 187 Google reviews, with 8 to 15 from the current month. Invisible practices have an average of 34 reviews, with the most recent dating back 2 to 6 months.

But volume alone isn’t enough. Review content makes a massive difference. Visible practices have reviews that mention specific services (“implant,” “cleaning,” “children’s orthodontics,” “whitening”) in over 40% of cases. Invisible practices have generic reviews (“great dentist,” “top practice”) in over 80% of cases.

AI systems literally extract keywords from reviews to build their understanding of what you do. A practice whose reviews mention “dental implant” 30 times will be associated with that query. A practice whose reviews only mention “very good” won’t be associated with anything specific.

Factor 2: Completeness of the Google Business Profile

Visible practices have a GBP completeness rate of 90% or higher: description of 700+ characters, all services listed individually, relevant primary and secondary categories, complete hours (including Saturday), booking link, over 20 photos, and at least one Google post from the last 15 days.

Invisible practices have an average completeness rate of 45%. The most common gaps: absent or under 200-character description (in 58% of audited practices), no services listed individually (in 44%), no Google post published in the last 3 months (in 72%), and fewer than 5 photos (in 36%).

Factor 3: Multi-Platform Presence and NAP Consistency

Visible practices are present on at minimum 5 platforms with strictly identical information: Google Business, website, Doctolib, PagesJaunes, and at least one professional directory (Dental Board, specialist directory, or network directory).

Invisible practices are present on 2 to 3 platforms on average, and 38% have at least one information inconsistency (different phone number, differently formatted address, contradictory hours).

This result confirms that AI systems systematically cross-reference data from multiple sources. Consistency is a trust signal; inconsistency is a signal of doubt.

Factor 4: Existence and Quality of the Website

Among the 50 audited practices, 34% have no website at all. None of these practices without a website appear in the visible group.

Among practices with a website, only 12% have structured data (Schema.org). Yet all 5 of the most visible practices have it. The correlation between structured data and AI visibility is the clearest in our study after reviews.

Factor 5: Recent Activity

Visible practices show activity signals within the last 30 days: new reviews, new Google post, added photo, content published on the site. Invisible practices often show a last activity dating back 3 to 12 months.

This factor is consistent with global trends: recently updated content receives significantly more AI citations than dated content.

What AI Systems Cite as Sources

A revealing aspect of our audit is the analysis of sources that AI systems cite when recommending a Parisian dentist.

On Perplexity (which displays its sources), the most frequent references are: Doctolib (cited in 68% of responses containing a dental recommendation), Google Maps / Google Reviews (54%), practice websites (31%), PagesJaunes (18%), and specialized health directories (12%).

On ChatGPT, responses don’t always explicitly cite their sources, but analysis of response content shows a strong influence from Google Reviews (the descriptions used by AI match Google reviews), Doctolib (specialties and availability), and practice websites (when they exist and are up to date).

On Google AI Overview, responses naturally draw from the Google index, with a clear preference for complete Google Business listings and websites with structured data.

The operational conclusion is clear: to be recommended by AI, a dental practice must be primarily optimized on Google Business, Doctolib, and its own website. Secondary directories reinforce the signal, but the three fundamental pillars remain Google, Doctolib, and the website.

Differences Between Neighborhoods

Our audit covers 12 arrondissements, allowing us to observe local variations.

The most competitive neighborhoods in terms of AI visibility are the 8th, 16th, and 17th. In these areas, practices that appear in AI recommendations have very strong profiles (150+ reviews, website with Schema, high GBP activity). The entry bar is high.

The neighborhoods with the most opportunity are the 10th, 12th, 14th, and 20th. In these areas, very few practices have optimized their presence for AI, meaning a practice that invests in GEO can quickly become the default recommendation in its neighborhood. The first to act gains the advantage.

The intermediate neighborhoods (5th, 9th, 11th, 15th) present a mix: a few well-positioned practices and a majority of invisible ones. The window of opportunity is still open, but narrowing.

Networks vs Independent Practices

An unexpected result from our study: practices belonging to networks (Dentylis, Dentego, etc.) don’t perform better than independents in AI visibility. In fact, their average GEO score is slightly lower (15/100 vs 20/100 for independents).

The likely explanation: the GBP listings of network practices are often standardized and less personalized. Descriptions are sometimes copied from one center to another, photos are generic, and reviews are diluted between the brand and the local center. AI systems, which look for specificity and authenticity, value an independent practice with a strong local identity more than a network center with a generic profile.

This is strategic information for independent practices: you have a natural advantage in GEO if you leverage your local roots.

What These Results Mean for Your Practice

If you’re a dentist in Paris, here are the three key takeaways.

First takeaway: the window of opportunity is wide open. With 62% of practices completely invisible and only 10% truly visible, the field is clear. A practice that invests today in a structured GEO strategy can become THE go-to recommendation in its neighborhood within a few months. AI systems tend to recommend the same sources repeatedly once they identify them as reliable — the first mover benefits from a cumulative advantage.

Second takeaway: the most impactful actions are accessible. The AI visibility factors we identified don’t require an advertising budget or advanced technical skills. Completing your Google Business listing, encouraging specific reviews, maintaining NAP consistency, and publishing regularly — these are actions any practice can take, with or without outside help.

Third takeaway: traditional search optimization isn’t enough. Among the audited practices, several have good traditional Google rankings (first page for “dentist [neighborhood]”) but a low GEO score. Ranking well on Google doesn’t automatically mean being recommended by AI. Recent studies show that only 14% of URLs cited by AI appear in Google’s top 10. These are two distinct channels that require complementary strategies.

Our Recommendations

Based on this study, here is our recommended action plan for any dental practice looking to improve its AI visibility.

Month 1 — Foundations (expected impact: +15 to +25 GEO points). Complete your Google Business listing to 100%. Launch a targeted review campaign encouraging patients to mention specific services. Align information across the 5 key platforms. Publish a first weekly Google Business post.

Month 2 — Reinforcement (expected impact: +10 to +15 additional GEO points). Implement Schema.org structured data on your website. Enrich your Doctolib profile (description, specialties, photos). Pre-populate the Q&A section of your Google listing. Produce fresh content on the site (FAQ, detailed service pages).

Month 3 — Acceleration (expected impact: +5 to +10 additional GEO points). Register your practice on missing directories. Maintain the pace of Google posts and new reviews. Test AI visibility and adjust the strategy based on results.

A practice that executes this plan can reasonably go from a score of 0-15 to a score of 40-55 in 90 days — enough to appear regularly in AI recommendations for its neighborhood.

---

Does this study concern you? Eddie Miller Agency provides a free individual GEO audit for dental practices. You get your personal GEO score out of 100, a comparison with your direct competitors, and a priority action plan. [Request my free GEO audit ->]

---

FAQ — AI Visibility Study for Dentists in Paris

Is this study representative of all dental practices in Paris?

Our sample of 50 practices covers 12 arrondissements and includes independent practices and network centers of varying sizes. The results are consistent with large-scale international studies (notably the SOCi 2026 Local Visibility Index on 350,000 businesses). The observed trends — low AI visibility across the board, concentration of visibility among a few well-optimized players — are found in all local markets.

My practice isn’t in the study. How can I find out my GEO score?

We offer a free individual GEO audit that uses the same methodology as this study. You receive your score out of 100, the breakdown by criterion, and priority actions. The process takes 48 hours.

Will the results of this study change quickly?

GEO scores are dynamic. A practice that optimizes its presence can see its score increase significantly in 2 to 3 months. Conversely, a practice that ceases all activity will see its score decline. We plan to publish a quarterly update of this study to track market evolution.

Why do AI systems recommend so few dental practices?

AI systems are designed to provide reliable recommendations. When the available information about a practice is incomplete, inconsistent, or outdated, the AI prefers not to cite it rather than risk an inaccurate recommendation. The confidence threshold required for an AI to recommend a business is higher than that required for a simple Google ranking.

---

You’re a dentist in Paris? Discover how we help dental practices get found ->