Pramaana Labs raised $27 million in seed funding, backed by Khosla Ventures, Accel, Boldcap, Nexus Venture Partners, Premji Invest, and Unbound.
The startup tackles a core problem in AI deployment: how to ensure reliability when errors could affect someone's health, freedom, or finances. The company uses formal verification—translating rules into mathematical proofs—to make AI systems deterministic and auditable.
Danny Werfel, former IRS commissioner, oversees Pramaana's tax law division. Co-founder and CEO Ranjan Rajagopalan describes the approach as formalizing rules so reasoning becomes deterministic. The method resembles France's CATALA project, which has been codifying legal language for years.
Pramaana layers a verification system on top of language models using LEAN, an open-source programming language for verifying mathematical proofs. The AI remains flexible and can answer natural language questions, but must also prove its reasoning.
The advisory board includes professors from IIT Delhi, IIT Madras, and UC Berkeley overseeing cybersecurity and drug discovery systems.
"Every domain where being wrong can cost someone their health, money, or freedom has rules," Rajagopalan told TechCrunch. "Now, those rules just need to be codified."
The company is betting that formalization will become standard practice for high-stakes AI applications.
