According to Forbes, former Google AI researchers Paulina Granarova, Kevin Roth, and Yannic Kilcher have raised a $41 million Series A for their legal tech startup DeepJudge, valuing the company at $300 million. The Zurich-based company, founded in 2020, builds customized indexes of law firms’ confidential archives that plug into existing AI models like ChatGPT. Their system helps lawyers search internal documents for answers to questions like “When have we done a deal like this before?” without manual document hunting. The funding round was led by Felicis with participation from existing investor Coatue, and comes as major US law firms including Freshfields and Holland & Knight have already signed on as customers. DeepJudge’s indexing process takes up to three weeks per client and specifically addresses the critical problem of AI hallucinations in legal work.
The real legal AI problem isn’t public data
Here’s the thing everyone misses about AI in law: ChatGPT knows the same stuff for every firm. The real competitive advantage isn’t in public information—it’s in the decades of confidential work product sitting in firm archives. DeepJudge’s CEO Paulina Granarova nailed it when she said a firm’s own data is their “true competitive power.” Think about it: Gunderson Dettmer has worked on nearly 2,000 financial transactions recently, with dozens of versions of the same documents. Before this, lawyers were manually opening each version to find what they needed. That‘s not just inefficient—it’s practically medieval in 2024.
Hallucinations are career-enders in law
When AI makes up case citations or contract terms, it’s not just embarrassing—it’s malpractice. Baker Donelson has already identified over 120 cases of hallucinations in court filings since 2023. That’s why DeepJudge’s approach of building contextual indexes is so smart. They’re essentially creating a verified map of a firm’s knowledge that AI models can reference, dramatically reducing made-up information. Coatue partner Caryn Marooney put it perfectly: “Legal of all things really can’t afford the hallucinations.” No kidding—your bar license depends on it.
The $8 billion legal AI race is heating up
DeepJudge’s $41 million might sound impressive, but they’re entering a brutally competitive space. Just last week, Harvey closed a $150 million round at an $8 billion valuation. Legora was raising at a $1.8 billion valuation. Basically, everyone and their mother is throwing money at legal AI right now. The timing makes sense though—law firms are desperate for solutions that actually work with their existing systems rather than requiring complete overhauls. DeepJudge’s plug-and-play approach with existing AI models could be their edge against the mega-funded competitors.
From PhD project to $300M company
What I love about this story is how organic the startup’s origin was. Three PhD students at ETH Zurich taking legal tech courses, their professor encouraging them, local Swiss law firms literally approaching them saying “if you can solve this, we’ll buy.” And they did solve it, and those firms did buy. Now they’re expanding aggressively into the US market with this new funding. Felicis partner James Detweiler said they’re “squarely entering act two about being a bit louder.” Translation: get ready for a lot more salespeople knocking on law firm doors. The question is whether their technical approach can scale fast enough to compete with the billions flowing to rivals.
