Partner Story ยท in conversation with Thomas Bavoil, founder of Klarr. This is a written interview submitted through Submit Your Story. The claims below are the founder's own, presented as they were given to us. No payment was involved.
Thomas Bavoil, founder of Klarr, Paris. Photo: Thomas Bavoil.
In July 2026, a fire hit a restaurant in Paris. It belonged to Michael Fink, a friend Thomas Bavoil has known since childhood. Fink lost his main working tool overnight and faced serious financial pressure.
"While helping him sort things out, I was shocked to see that some booking systems can cost a restaurant hundreds of euros," Bavoil says.
Bavoil is a web developer and SEO specialist who runs EDIREF, a Paris web studio in business since 2008. Over the years, many of the studio's clients were independent restaurants, and the same problems kept coming back. Fink had always been building things. Bavoil offered to help, and Klarr grew out of that: one tool that helps an independent restaurant get found online, above all by AI assistants, with a website, commission-free booking and publishing in the same place.
The problem: invisible where diners now decide
A restaurant's Google listing is often incomplete or out of date. Its opening hours differ from one platform to the next. Many independent places have no real website, and booking platforms take a commission on every cover.
Meanwhile, more and more people simply ask an AI assistant where to eat tonight. "If a restaurant isn't in that answer, or appears with the wrong phone number or hours, it loses customers without ever knowing it," Bavoil says.
Big chains have marketing teams to handle this. "A family bistro doesn't."
What six AI assistants said about Paris
To see how big the problem is, Klarr ran a test. On 4 October 2026, the team asked ChatGPT, Claude, Gemini, Grok, Le Chat and Perplexity for the best restaurant in each of the 20 arrondissements of Paris, then checked the first restaurant named in every answer.
The 120 answers named 82 different restaurants first, and in no arrondissement did all six assistants agree. Seventeen answers sent the diner to the wrong place: seven recommended a restaurant that had closed, and ten gave the wrong address or the wrong district. Klarr published the results in a study on its site, and says there openly that it is not a neutral party, since it sells the tool that measures this.
Where the AI sits
AI runs through all of Klarr, in three jobs.
The first is the audit. Klarr asks the assistants people actually use (ChatGPT, Gemini, Perplexity, Claude, Le Chat, DeepSeek and Doubao) the kinds of questions a diner would ask, in several languages. It then scores the restaurant on two things. Visibility: does the AI mention the restaurant at all, and how high? Accuracy: are the address, hours, phone number and concept right?
The second is content. Claude writes the restaurant's posts and its website copy, and keeps its information consistent everywhere, "because AI assistants trust restaurants whose information matches across sources."
The third is publishing. Since October 2026, following Meta's validation, Klarr posts to Facebook, Instagram and the restaurant's Google listing. For RedNote (Xiaohongshu), which matters for Chinese-speaking visitors in Paris, Klarr prepares the whole post and the restaurateur publishes it in one tap.
The product runs mainly on Claude by Anthropic, with Mistral AI, a French and European model, as a backup.
One flat price, no commission per cover
Most tools solve one piece of this. An agency builds a website, a booking platform takes reservations and charges per cover, a social media tool schedules posts. Klarr puts it all in one place for a flat monthly price.
The prices are on the site: 37.50 euros a month for the visibility plan, 29 euros a month for bookings, or 59 euros for both, excluding VAT, per restaurant and with no contract. Klarr's own comparison shows why that matters to an owner: at 400 covers a month, a platform charging 2 euros per cover takes 800 euros.
Klarr also starts from the question that classic SEO tools do not measure: what do AI assistants say about your restaurant? "We chose to test several AI engines rather than one, because diners don't all use the same assistant, and a restaurant can rank first in one and not exist in another," Bavoil says.
What the first audits found
Klarr is live and in use with independent restaurants in Paris. Each new restaurant gets an AI visibility audit when it signs up, and Bavoil calls the findings remarkably consistent. One restaurant had three different phone numbers online, opening hours that did not match between platforms, and a Google listing pointing to a third-party page instead of its own website. Those are exactly the errors that make an assistant give a wrong answer or skip a restaurant, and they are the first things Klarr fixes.
Restaurant owners already using it describe the everyday side.
"With two restaurants, I don't have time to juggle ten tools. With Klarr, bookings, reviews and my Google listing are in one place, and I pay no commission," says An, owner of Kokodak and Joayo 13.
"At TAN, we also have karaoke: private hire and deposits used to be a headache. With Klarr, customers book their room, pay their deposit online, and everything comes straight to us," says Xuanmin, founder of TAN.
"During service, I see the whole dining room table by table, and requests come straight in. No more booking book, no more calling every customer back," says Huijun Yan, manager of Prost.
The lesson: show the audit, skip the pitch
"The obstacle we still face is trust and time," Bavoil says. Restaurant owners are overwhelmed, and many have been burned by tools that promised visibility and delivered invoices. Explaining that AI visibility is real, and already costing them customers, takes patience.
What works is showing owners their own audit, where they can read what ChatGPT or Gemini says about their restaurant. That, Bavoil says, works far better than any pitch.
Where it stands today
"We're early, and we'll share numbers as the Paris test grows," Bavoil says. One question is still open: how to reach restaurant owners at scale. The team says it would welcome advice on that.
What comes next
The plan is to bring more independent restaurants on board in Paris, then expand across France, and to keep improving the audit so that it tracks changes over time.
Klarr is looking for restaurants willing to test it, for partners in the hospitality world such as schools, federations and suppliers, and for advice on reaching small business owners. The product, the prices and a free AI test are at klarr.net.
---
This is a Partner Story: a written interview with a founder building with AI, submitted through Submit Your Story and published free of charge. Statements about the company's product, users and results are its own. Building something with AI? Tell us about it.
Would you try this product?
