As enterprises struggle to turn AI pilots into functional parts of their business, reliability has taken center stage. A new startup hopes to solve this problem by leveraging the tools of mathematical modeling, combining one of computer science’s most reliable systems with one of its most chaotic.
On Wednesday, Pramaana Labs announced $27 million in seed funding led by Khosla Ventures, with participation from Accel, BoldCap, Nexus Venture Partners, Premji Invest and Unbound.
Pramaana will focus on highly sensitive industries such as law, drug discovery and tax preparation – where errors can be costly and reliability is at a premium. The development of artificial intelligence in these systems will require stronger protection against hallucinations and errors than we have today. But as Pramaana co-founder and CEO Ranjan Rajagopalan sees it, they’re also uniquely suited to formalization.
“It’s like math in the sense that you have a lot of rules to follow,” Rajagopalan told TechCrunch, describing the rules of the tax code. “Once you have a coded version of it, reasoning on top of it starts to become deterministic.”
Pramaana’s system still works with a conventional LLM, giving it the flexibility to answer natural language questions and tackle complex problems that conventional computers can’t handle. But there is a deterministic layer on top of this LLM that ensures that the LLM project will be checked.
This combination of LLM engine with deterministic verification is a popular setup. Pramaana’s unique approach is to use formal verification tools — drawing from the open source LEAN programming language used to verify mathematical proofs. There is real precedent for much of this work. Rajagopalan points to France’s CATALA projectwhich formalizes much of the country’s tax and benefits system into executable code.
For each use case, Pramaana will create its own formal LEAN verification system overseen by industry experts. For tax law, the company is working with former IRS commissioner Danny Werfel, while professors from IIT Delhi, IIT Madras and UC Berkeley are overseeing the cybersecurity and drug discovery system.
“The world’s hardest problems aren’t unsolvable. They’re unsolvable,” says Rajagopalan. “Every field where a mistake can cost someone their health, money or freedom has rules.”
Now, these rules just need to be encoded.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
