GEO Barometer Paris Q1 2026: Which Businesses Do AI Assistants Recommend Most?

Every quarter, Eddie Miller Agency tests the AI visibility of Parisian businesses on ChatGPT, Perplexity, and Google AI Overview. The goal: track market evolution, identify trends, and provide local professionals with a reliable benchmark for their visibility across AI assistants.

This first barometer covers three sectors: dental practices, restaurants, and hotels. For each sector, we tested a sample of businesses across several Parisian districts using typical queries.

Here are the results for the first quarter of 2026.

Barometer Methodology

For each sector, we submitted between 20 and 30 queries phrased the way a real user would, across three platforms: ChatGPT, Perplexity, and Google AI Overview. The queries covered different districts, different specialties or cuisine types, and different levels of intent (general search, comparison, urgency, name-based queries).

Results are aggregated into three indicators: the mention rate (percentage of queries where at least one business from the sector is cited), the accuracy rate (percentage of factually correct information in responses), and concentration (how many different businesses are cited — many or always the same ones).

Dental Sector: AI Visibility Concentrated Among Very Few Players

The finding is clear. Across the 30 dental queries tested in 10 Parisian districts, here is what we observed.

Mention rate by platform. Perplexity recommends at least one specific dental practice in 73% of cases. It is the most generous platform for dental recommendations, largely because it relies heavily on Doctolib, which is a very rich source for this sector in France. Google AI Overview mentions specific results in about 45% of cases, mainly when a Map Pack is triggered and the query is sufficiently local. ChatGPT recommends named practices in only 38% of cases. For the remaining queries, it gives generic advice (“check Doctolib”) without naming anyone.

Concentration is extreme. In the districts tested, the same 2 to 3 practices come up repeatedly. This means a small number of practices capture nearly all AI recommendations, while the vast majority are never cited. This is consistent with international data: the SOCi 2026 report showed that only 1.2% of local businesses are recommended by ChatGPT.

Factors observed among dominant practices. Unsurprisingly, these are the same ones identified in our detailed study on 50 dental practices: high volume of recent reviews containing service-related keywords, a complete Google Business profile, presence on Doctolib with a comprehensive profile, and for some, a website with structured data.

Emerging trend this quarter. We observe a slight increase in mentions of dental networks (Dentego, Dentylis) by ChatGPT on generic queries, likely because these brands have more developed websites with editorial content. On Perplexity, however, independent practices with strong Doctolib profiles remain better positioned.

Restaurant Sector: Fierce Competition but Real Opportunities

The Parisian restaurant scene is a different playing field from healthcare. More options, more volatility, and a much stronger role for review and booking platforms.

Mention rate by platform. Perplexity recommends specific restaurants in 89% of cases — the highest rate of all three sectors studied. Restaurants are the category best covered by Perplexity, which extensively cites TripAdvisor, TheFork, and food blogs. ChatGPT recommends named restaurants in 67% of cases, significantly more than for dentists. Conversational AI naturally excels at restaurant recommendations since it is a very frequent use case. Google AI Overview is more restrained, mentioning specific restaurants in about 35% of cases, most often when an enriched Map Pack is triggered.

Concentration is less extreme than for dental. AI assistants recommend a broader range of restaurants — typically 5 to 8 different names for the same query asked 10 times, compared to 2 to 3 for dentists. This is explained by the diversity of options and the number of available sources (blogs, guides, reviews across multiple platforms).

Dominant sources for Parisian restaurants. TripAdvisor is the most cited source by Perplexity for restaurants in Paris, followed by Google Reviews and TheFork. ChatGPT relies more on Google Business, food blogs, and “best restaurants” lists published by media outlets. An interesting observation: restaurants mentioned in “top 10 [cuisine type] in Paris” articles published by recognized media (TimeOut, Le Fooding, Le Figaro) have a marked advantage on ChatGPT.

Emerging trend this quarter. Long conversational queries (“romantic restaurant with a terrace in the Marais, French cuisine, average budget 40-50 per person”) produce much more precise and differentiated recommendations than short queries (“restaurant Paris 3rd”). Businesses whose online profiles detail the ambiance, price range, and special occasions are better positioned for these high-intent queries.

Hotel Sector: The Most Mature in AI Visibility

The Parisian hotel industry is the most advanced sector in terms of AI visibility, primarily because booking platforms (Booking.com, Expedia, Google Hotels) provide very rich structured data that AI assistants exploit directly.

Mention rate by platform. Perplexity recommends specific hotels in 92% of cases — the highest rate in the entire barometer. ChatGPT recommends named hotels in 78% of cases. Google AI Overview mentions hotels in about 55% of cases, often with price ranges and booking links.

The dynamics are different. Unlike dental and restaurant sectors, the Parisian hotel industry is already heavily structured online. Major booking platforms (Booking, Expedia, Hotels.com) serve as massive data intermediaries. AI assistants draw directly from these platforms to formulate their recommendations.

The challenge for independent hoteliers. Hotel chains and hotels listed on all booking platforms dominate AI recommendations. Independent hotels or boutique hotels that are only present on their own website and Booking have a visibility deficit on ChatGPT and Perplexity. Their challenge is to diversify their presence: Google Business, TripAdvisor, travel guides, and specialized blogs.

Emerging trend this quarter. Experience-related queries (“charming hotel with a view of Paris,” “design boutique hotel in the Marais”) favor businesses whose reviews explicitly mention these experience elements. Hotels that encourage detailed reviews describing the atmosphere and experience (not just cleanliness and value for money) are better positioned for high-value queries.

AI Platforms: Who Recommends What and How

This barometer also allows us to refine our understanding of each AI platform’s behavior.

ChatGPT remains the most selective platform. It recommends less often and less consistently. But when it does recommend, it does so with a conversational tone that strongly influences the customer’s decision. Studies show that ChatGPT holds about 80% market share among AI chatbots — making it the hardest platform to appear on, but also the most impactful.

Perplexity is the most generous platform for local recommendations. Its citation-first approach (all sources are visible) makes it a high-trust channel. For professionals, Perplexity is also the most predictable platform: recommendations change little from one test to another, meaning once you are visible, you tend to stay visible.

Google AI Overview remains the most conservative and the most accurate. It only recommends when it is confident in its data, which explains its lower mention rates. But its potential impact is enormous: it reaches all Google users without them needing to go to another platform.

Districts: Where the Opportunities Are

A fascinating aspect of the barometer is the disparity between districts.

The most competitive districts (all sectors combined) are the 1st, 6th, 8th, and 16th. In these areas, AI assistants have a wide selection of well-referenced businesses, and the bar for being recommended is high.

The districts with the most opportunities are the 10th, 12th, 13th, 14th, 19th, and 20th. In these areas, very few businesses have optimized their AI presence. A dental practice, restaurant, or hotel that invests in GEO in these districts can become the default recommendation within a few months.

The intermediate districts (5th, 9th, 11th, 15th, 17th, 18th) present a mix of well-established leaders and large vacant spaces. The opportunity is real, but time is pressing — each quarter, new players optimize their presence.

What This Barometer Reveals

Three major takeaways emerge from this first barometer.

First takeaway: AI visibility is a zero-sum game at the local level. When an AI recommends 3 businesses in a district, the other 50 do not exist. Every recommendation gained by a competitor is a recommendation lost for you. Inaction is not neutral — it is costly.

Second takeaway: all three sectors follow the same fundamental logic. The factors that determine AI visibility are remarkably consistent across sectors: information consistency, recent and specific reviews, complete profiles on key platforms, and structured data on the website. Only the source platforms vary (Doctolib for dental, TripAdvisor/TheFork for restaurants, Booking for hotels).

Third takeaway: the window narrows each quarter. As professionals become aware of the importance of GEO, positions solidify. The first to optimize their presence gain a cumulative advantage that latecomers will struggle to close. This first barometer serves as a reference point — the next one (Q2 2026) will show the scale of the movement.

The Next Barometer

The GEO Barometer Paris Q2 2026 will be published in July 2026 with an expanded sample, additional queries, and tracking of how previously analyzed businesses have evolved. If you would like your business to be included in the next barometer, contact us via the free audit below.

---

Where does your business stand in this landscape? Eddie Miller Agency provides a free individual GEO audit for Parisian dental practices, restaurants, and hotels. You receive your AI visibility score, your positioning relative to competitors, and priority actions. [Request my free GEO audit →]

---

FAQ — GEO Barometer Paris

How often is this barometer published?

The barometer is published every quarter (January, April, July, October). Each edition includes the most recent data, changes from the previous quarter, and newly observed trends.

Can my business be included in the barometer?

Yes. If you are a dental practice, restaurant, or hotel in Paris, you can request to be included in the next barometer through our free GEO audit. We will test your AI visibility using the same methodology as the barometer.

Do the barometer results change quickly from one quarter to the next?

AI positions are more dynamic than traditional Google rankings. A business that actively optimizes its presence can see significant changes within 60 to 90 days. Conversely, a business that stops being active online (no new reviews, no posts) can lose its visibility within a few months. The quarterly barometer captures these movements.

Are the barometer data applicable to cities other than Paris?

The fundamental principles (importance of reviews, NAP consistency, structured data) are universal. However, source platforms and competitiveness vary by city. This barometer is specifically calibrated for the Parisian market.