Consint.AI Raises ₹22 Crore to Turn Its Claims-Fraud Engine Into a Foundation Model

An AI built to watch where the money leaks

India routes a fast-growing share of its healthcare and insurance rupees through digital claims, and a stubborn slice of those claims never should have been paid. Consint.AI, a Bengaluru-based deep-tech startup working on exactly that problem, has raised ₹22 crore — roughly $2.3 million — in a Series A round to expand overseas and, more ambitiously, to build a purpose-made artificial-intelligence model for spotting fraud, waste and abuse across healthcare, insurance, banking and financial services.

The company's pitch rests on a track record it is happy to quantify. By its own account, the platform has already assessed more than 100 million transactions and flagged over ₹1,000 crore worth of fraudulent or wasteful activity, drawing on a stack of more than 500 machine-learning models, several fine-tuned large language models, and a library of over 500 digitised clinical protocols that let it reason about whether a given medical claim actually makes sense.

What the platform actually does

"Fraud, waste and abuse" — FWA, in the industry's shorthand — is a quieter leak than outright theft. It is the duplicate claim, the up-coded procedure, the treatment that does not match the diagnosis, the bill that is technically permissible but clinically indefensible. For insurers, third-party administrators, hospitals and government health schemes, catching it after the money has gone out is slow and adversarial; catching it before payout is where the savings live.

Consint.AI positions itself in that pre-payment window. Instead of a single fraud score, it runs a mesh of specialised models against each claim, cross-checking the clinical logic of a treatment against its own protocol library and the claimant's history. Fine-tuned language models let the system read the unstructured notes — discharge summaries, physician remarks — that rule-based engines have historically ignored. The output is meant to slot into an insurer's existing claims-processing workflow, improving both the accuracy of risk assessment and the speed of legitimate settlements.

The round and its backers

The Series A was led by BIG Global Investment JSC, with participation from Equanimity Ventures Trust II and Seafund's Venture India Scheme I. Founded by Ashish Chaturvedi, Consint.AI describes itself as a deep-tech company transforming healthcare and insurance risk management through AI, serving insurers, healthcare providers and government health programmes.

The capital, the company says, will go toward accelerating international expansion, strengthening enterprise delivery, and — the line that matters most for its long-term story — developing a foundational AI model dedicated to fraud, waste and abuse detection across healthcare, insurance, banking and financial services.

From point tool to foundation model

That last ambition is the interesting one. Most Indian applied-AI startups sell a workflow: a tool that plugs into a customer's stack and does one job better. Building a "foundation model" for FWA is a bet that fraud patterns across domains — a padded hospital bill, a staged motor claim, a laundered banking transaction — share enough underlying structure that a single large model, trained on the right data, can generalise across them.

It is a defensible bet precisely because Consint.AI already sits on the scarce ingredient such a model needs: labelled, real-world transaction data at scale, tied to outcomes. The 100-million-transaction figure is not just a marketing number; it is the training corpus.

Why it matters for India's AI stack

India's AI conversation has spent the past two years fixated on frontier models and sovereign compute. The less glamorous — and arguably more bankable — story is applied AI aimed at concrete leakage in large systems, and few systems leak like insurance and public health spending. A model that measurably reduces improper payouts has a return on investment that does not require any leap of faith about artificial general intelligence.

Consint.AI's raise is modest by the standards of India's headline AI rounds, but it is squarely in the category investors have gravitated toward through 2026: deep-tech with a paying enterprise customer, a defensible data moat, and a problem worth billions if solved even partially. Whether the company can turn a working detection engine into a genuine foundation model is the open question — and the reason the next 18 months will matter more than this cheque.

Sources

• Inc42 — "Consint.AI Bags ₹22 Cr To Expand Globally, Build AI Model For Risk And Fraud Detection"
• Entrackr — "Healthcare AI startup Consint.AI raises Rs 22 Cr in Series A round"
• SiliconIndia — "Consint.AI Raises Rs 22 Crore to Expand AI Healthcare Platform"
• DigitalHealthNews — "Consint.AI Secures INR 22 Cr Series A to Scale AI-Led Healthcare Fraud Detection Worldwide"