Skip to main content

Pramaana Labs

Category: AI Infrastructure

Pramaana Labs builds a formal verification layer for AI, making outputs provably correct for high-stakes regulated industries like tax, law, and healthcare. Pramaana Labs was founded in 2025. The company is led by Ranjan Rajagopalan. Based in Bengaluru, India (HQ) & Palo Alto/San Francisco, CA, USA. Team size: 11-50. Total funding raised: $27M. Latest round: Seed. Key investors include Khosla Ventures (lead), Accel, BoldCap, Nexus Venture Partners, Premji Invest, Unbound.

Founded
2025
Headquarters
Bengaluru, India (HQ) & Palo Alto/San Francisco, CA, USA
Team size
11-50
Total funding
$27M

Value proposition

Pramaana builds a verification layer for AI that translates complex domain knowledge (tax codes, clinical guidelines, legal statutes) into machine-checkable formal representations, returning proof artifacts that domain experts can inspect — ensuring AI outputs are provably correct, not just probabilistically plausible.

Products and solutions

Pramaana Stack: Domain Formalizer (converts regulatory/statutory text into formal representations), Prover & Solver engines (search solution space for proofs), Proof Artifact Generator (returns traceable, machine-checkable proofs). Named components: Leibniz, Panini, Ramanujan, Hardy. First vertical focus: Tax & statutory reasoning.

Unique value

The only AI verification platform using formal verification (LEAN programming language) to deliver machine-checkable proofs of correctness for regulated industries, rather than confidence scores or probabilistic outputs.

Target customer

Enterprises in regulated industries: tax compliance firms, legal departments, healthcare organizations, drug discovery companies, financial compliance teams, and government/public sector agencies.

Industries served

Tax & statutory reasoning, Legal & regulatory compliance, Healthcare & clinical safety, Drug discovery, Cybersecurity, Financial compliance

Technology advantage

Combines LLM flexibility with deterministic formal verification using the open-source LEAN programming language (used to verify mathematical proofs). Auto-formalization converts natural-language rules into machine-checkable specifications. Inspired by France's CATALA project for formalizing tax/benefit systems. Domain experts oversee each vertical's verification system.

How they differentiate

Unlike conventional AI guardrails that provide confidence scores or citations, Pramaana delivers mathematical proofs of correctness. The system either returns a machine-checkable proof or refuses to answer if a proof cannot be established. This is fundamentally different from probabilistic approaches — it's a compiler for mission-critical AI, not a filter.

Main competitors

Axiom Math (formal verification for code), Amazon Bedrock Guardrails with Automated Reasoning, Theorem AI (code verification)

Key partnerships

Former IRS Commissioner Danny Werfel (tax formalisation advisor), Professors from IIT Delhi, IIT Madras, UC Berkeley (research ecosystem), Stanford University's Centaur Lab (collaboration)

Major milestones

Founded September 2025, Raised $27M seed round (June 2026) — one of the largest seed rounds in AI, RuleArena benchmark (ACL 2025) cited as evidence of the problem space, Domain Project 1729 announced on website

Market positioning

Frontier AI lab focused on the verification layer for high-stakes AI. Positioned at the intersection of formal verification (traditionally used in mathematics/software verification) and large language models. Targets the growing enterprise need for reliable, auditable AI in regulated industries.

Geographic focus

Global (US-headquartered operations with Indian R&D hub)

About Ranjan Rajagopalan

Ex-Google (led Google Maps Moderation); Co-founded Astra; Software Engineer at Graviton Research Capital LLP; Dual Degree in Computer Science from IIT Madras (2012-2017)

Official website: