Bravi raises $1.8M seed to put an AI product expert in every field rep’s pocket

Bravi, founded in Cholet in 2025 and a graduate of Y Combinator’s F25 batch, has closed a first round of $1.8M from American and French funds, European and American business angels and Y Combinator itself, to keep building its AI layer for home-improvement manufacturers and to open a commercial front in the United States.

A first round closed between Maine-et-Loire and Brooklyn

The round is $1.8M, and it is the company’s first, announced on 15 September 2026. The cap table mixes American and French funds with business angels on both sides of the Atlantic, plus Y Combinator, which took Bravi into its Fall 2025 batch for three months in San Francisco. No lead investor has been named publicly, and no valuation has been disclosed.

The money goes to two places at once: new product for the customers Bravi already has in France, and a commercial beachhead in the United States. That second half is unusual at this size. Most French companies at $1.8M are still arguing about whether to hire a second salesperson in Lyon.

Co-founder Anas Bouassami describes the batch as intense, and says it did two things at once: it put investors in front of the team, and it gave Bravi far more visibility back in France. Read that twice. A French founder flew to California and came back more visible at home.

Two returning founders and the least glamorous buyer in software

Bravi was founded in 2025 by Anas Bouassami and Pierre-Habté Nouvellon, who previously built Snipfeed together, acquired in 2024. Both list Forbes 30 Under 30 US on their profiles. This is a second company, not a first one, and it shows in the choice of market.

The product is an AI layer over the data a window, door or joinery manufacturer already owns and cannot find: technical documentation, product catalogues, ERP records, pricing grids. On top of that sit a quoting configurator, a conversational agent for technical questions, and an AI receptionist that answers calls around the clock. The company’s own line is “a technical product expert in every field rep’s pocket”, and roughly twenty companies in France and the United States are using it, from SMEs to large groups.

The team is around eleven people, split across Cholet, Paris, San Francisco and Brooklyn. You can find the full record on the Bravi profile and the wider category on our construction tech hub.

A 26% growth rate and a sector that never bought software

The European market for AI in construction was worth $1.43bn in 2025 and is forecast to grow at a 25.92% CAGR to 2034, according to MarketDataForecast’s July 2026 report. The growth rate is not the interesting part. The interesting part is the denominator: a sector where the quote still leaves as a PDF a human assembled by hand from three catalogues and a price list that changed last week.

For scale, our own fundraising data puts this round squarely in the current European vertical AI seed band. Logibot raised a €1.4M seed on 17 September 2026 for logistics automation; Complir raised an $11M seed led by General Catalyst for product compliance. Bravi sits at the lower end, which is exactly what a first institutional round after YC usually looks like, and it buys eighteen months rather than a land grab.

Why American money showed up for a French joinery tool

To understand why US funds sit on a $1.8M French cap table, read what Sequoia Capital published on 5 March 2026. In “Services: the new software”, partner Julien Bek argues that the next generation of AI companies will not sell tools to professionals but sell the work itself, and he puts a number on the prize: for every dollar spent on software, six are spent on services. His frame is copilots versus autopilots. A copilot helps the estimator write the quote. An autopilot produces the quote, and answers the phone at two in the morning.

Bravi is an autopilot dressed as a software company. The AI receptionist replaces a call-answering service. The quoting configurator replaces hours of estimator time. The technical agent replaces the one person in the factory everyone phones when a customer asks about a U-value. None of that comes out of a software budget, because most fenestration manufacturers barely have one. It comes out of payroll and outsourcing, which is precisely the entry point Bek describes: start where the customer already pays someone else to do the work, then move into the work they do in-house as the models improve.

That is the American thesis, and it explains the American cheque. It also explains why the market figure above understates the opportunity. $1.43bn is the software market for AI in European construction; the services budget it can absorb is a multiple of that, and nobody has sized it credibly for this trade yet.

What a Cholet cap table says about 2026

Two more things are worth watching. For a decade the story was that a French company went to YC and stayed in San Francisco. Bravi did the batch, took American money, and kept the engineering and the first twenty customers in Maine-et-Loire. The good news is that this is now a normal shape rather than a heroic one: the accelerator is a distribution channel, not an emigration route.

The second thing to watch is the buyer. European vertical AI has spent two years selling to companies that already had a software budget. Bravi is selling to fenestration manufacturers, which mostly did not. That is a harder sale and a much better moat, because whoever wins that relationship owns the catalogue, the pricing logic and the quoting workflow of an entire trade.

Watch the next twelve months for whether the US customers convert at the same rate as the French ones. That number, more than the round, tells us whether the fenestration trade is a French market with an American extension or a genuinely transatlantic one. Keep building.