Pramaana Labs Lands $27M to Make AI More Reliable

Artificial intelligence is becoming a bigger part of business operations, but many companies still struggle to trust AI with critical tasks. While AI models are powerful, they can also produce incorrect or misleading responses, commonly known as hallucinations. A new startup, Pramaana Labs, believes it has found a solution by combining advanced AI with mathematical verification.

On Wednesday, the company announced that it has secured $27 million in seed funding in a round led by Khosla Ventures. The investment also attracted support from Accel, Boldcap, Nexus Venture Partners, Premji Invest, and Unbound, giving the startup fresh capital to develop its technology.

Building AI That Businesses Can Trust

Pramaana Labs is targeting industries where accuracy is essential and even small mistakes can have serious consequences. These include sectors such as tax preparation, legal services, and drug discovery, where businesses need dependable AI systems before they can fully integrate them into daily operations.

The company aims to reduce AI errors by adding a verification layer that checks whether an AI model’s output follows established rules. This approach is designed to make AI systems more dependable in environments where precision matters most.

According to Pramaana co-founder and CEO Ranjan Rajagopalan, many professional fields already operate under detailed rulebooks. Tax law, for example, follows clearly defined regulations that can be translated into structured logic.

He explained that once these rules are properly codified, AI reasoning becomes much more predictable and deterministic instead of relying solely on probabilities.

Combining Large Language Models With Formal Verification

Rather than replacing traditional large language models (LLMs), Pramaana enhances them. The company’s system still uses conventional LLMs to understand natural language, answer questions, and solve complex problems. However, every response is checked through a deterministic verification layer before it can be trusted.

This hybrid approach—combining generative AI with deterministic validation—is becoming increasingly popular. What sets Pramaana apart is its use of formal verification, a technique commonly used in computer science to mathematically prove that software behaves correctly.

The startup relies on the open-source LEAN programming language, which is widely known for verifying mathematical proofs. Rajagopalan pointed to France’s CATALA project as an example of how tax and benefit systems can be transformed into executable code using similar principles.

Experts Help Shape Every System

Instead of creating a single verification model for every industry, Pramaana plans to develop separate LEAN-based verification frameworks tailored to each specific field.

These systems will be built under the guidance of subject-matter experts. For its tax-related platform, the startup is working alongside former IRS Commissioner Danny Werfel. Meanwhile, professors from IIT Delhi, IIT Madras, and UC Berkeley are contributing expertise to projects focused on cybersecurity and drug discovery.

Rajagopalan believes many of the world’s most difficult challenges are not impossible to solve—they simply haven’t been formalized into clear, machine-readable rules.

“Our hardest problems are not unsolvable. They are unformalized,” he said, adding that industries involving health, money, and personal freedom all depend on strict rules that can eventually be translated into reliable AI systems.

With its new funding, Pramaana Labs plans to continue transforming these complex rule-based systems into dependable AI tools, helping enterprises move beyond experimental AI projects and toward real-world deployment with greater confidence.

Also read : India Temporarily Blocks Telegram Over NEET Exam Fraud Fears