The Franco-American start-up completes a lifting of $ 15 million, including 2 million debt to, among other things, attack the European market more.
The future of intellectual property will also pass through AI. And Deepip understood this. The young Franco-American start-up founded three years ago is developing an AI system aimed at automating, in part, the writing of patents. The goal? Getting time and money for firms specializing in intellectual property and large groups.
Deepip was founded within Key Technology, a company specializing in data to improve the quality of the models. Since its creation, Deepip has targeted intellectual property, an area that had the double advantage of having specialized and spared by automation. The sector is “massively based on writing, and the human cognitive limits create major ineffectives there. There are a lot of patents which are poorly written, insufficient quality, and which do not effectively protect underlying inventions”, explains François-Xavier Leduc, CEO and co-founder of the company.
Writing patents and letters of rejection
Deepip has already developed two additional products. The first, marketed about a year ago, is an intelligent editorial assistant who joins the working environment of lawyers directly. The co -pilot fits into Microsoft Word and saves up to 50% of the writing time. A time saving but also financial for large groups. “In the United States, an hour of lawyer costs 500 dollars, against 300 euros in Europe. So, at this level, we are talking about concrete savings: between 2,000 and 6,000 euros for the drafting of a patent,” said François-Xavier Leduc.
The second product, launched four months ago, is dedicated to prosecution. “Once you have filed a patent, the regulatory bodies will analyze it and indicate to you if it is not valid for various reasons. Lawyers must then send these objections,” said the CEO. The solution allows, again, to save 30 to 40% of the writing time for rejection letters.
Technically, the Deepip solution is based on a complex architecture combining several models of generative artificial intelligence. “We use two models of fundamental language on different stages of writing. These are supplemented by quality models that evaluate each generation and relaunch the process until the result corresponds to our criteria,” explains François-Xavier Leduc.
15 million to accelerate Go-to-Market
To advance more quickly to the market, Deepip has just raised $ 13 million in series A plus 2 million debt from renowned investors such as Resonance, Headline (already present at Mistral AI), Serena Capital (Dataiku) and Balderton Capital. New funds that will mainly serve to accelerate the marketing of solutions and the development of new products, always in intellectual property. The company is currently strengthening its sales teams on both sides of the Atlantic. “We have just recruited three people in the United States and two others are being integrated. We also expand our team in France to adopt a more proactive and structured approach on the market,” said the CEO.
Technological development is the other major priority. “Our plan is to double our technical team this quarter,” specifies the manager. And to add: “Within three months, we will still double our workforce both on the product and on marketing.” With 80% of its turnover currently achieved across the Atlantic, the company, which has already won over the largest American firms specializing in intellectual property, now intends to strengthen its presence in Europe, a market that would represent around 30% of the global value of the sector. And the figures are already there. In just three months, “the company achieved 50% of its growth recorded throughout the year 2024”.
