Pramaana Labs has secured a $27 million seed round led by Khosla Ventures, with backing from Accel, Boldcap, Nexus Venture Partners, Premji Invest, and Unbound. The startup is building a new approach to AI reliability by combining large language models with formal verification methods from mathematics and computer science.
The company is targeting high-stakes fields such as law, drug discovery, and tax preparation, where accuracy matters as much as speed. Its system uses a conventional LLM for flexible language understanding, then adds a deterministic verification layer to check whether the output follows the required rules.
Co-founder and CEO Ranjan Rajagopalan says the idea is to turn complex rule-based domains into codified systems that AI can reason over more reliably. Pramaana is using the open-source LEAN programming language as part of that verification framework, a method designed to support mathematical proof checking and structured reasoning.
To adapt the model for each sector, the company plans to build domain-specific verification systems guided by experts. For tax law, it is working with former IRS commissioner Danny Werfel, while professors from IIT Delhi, IIT Madras, and UC Berkeley are advising on cybersecurity and drug discovery applications.
As AI moves deeper into regulated and precision-driven industries, approaches that combine generative capability with formal checks may help define the next phase of trustworthy automation. That shift could shape how intelligent systems are built across critical sectors in the years ahead.