Pramaana Labs raises $27M seed round led by Khosla Ventures

In short: Pramaana Labs says it raised $27 million to build AI tools that can double-check their answers in high-stakes fields.

What happened

Pramaana Labs announced a $27 million seed funding round led by Khosla Ventures. Other investors included Accel, Boldcap, Nexus Venture Partners, Premji Invest, and Unbound.

The company says it is focusing on areas where mistakes can be expensive or harmful, including law, drug discovery, and tax preparation. These are domains with lots of strict rules, which Pramaana believes can be written down in a way a computer can follow.

Pramaana’s approach combines a standard large language model, or LLM (the kind of AI that writes text like a chatbot), with what it calls a “deterministic” checking layer. Deterministic means the same input should always lead to the same result, like a calculator. The goal is to reduce “hallucinations,” which is when an AI confidently makes up something that is not true.

The company plans to use tools from formal verification, which is a method of proving a system follows rules using math, similar to how you can prove a math proof step by step. It is drawing on the open-source LEAN programming language, which is used to verify mathematical proofs. Pramaana says it will build separate rule-checking systems for each domain and have experts oversee them, including former IRS commissioner Danny Werfel for tax work.

Why it matters

Many people now meet AI through customer support, work tools, and filing or legal services. If AI is going to be used for decisions about money, health, or legal rights, it needs guardrails that work like a spellchecker for logic, not just grammar.

Source: TechCrunch AI