The launch points to a new phase of payment optimisation, where AI is used to improve transaction outcomes as well as detect risk.
Razorpay introduced Vulcan, an AI foundation model developed for payment use cases, at Global Fintech Fest 2026. The company said the model was trained on payment data at large scale and developed with technology support from NVIDIA and Amazon Web Services.
The product is designed to help payment systems make better decisions around transaction success, risk and payment behaviour. Razorpay has reported an 8–10% improvement in payment success rates during beta testing; this is a company-reported result and should be read in that context.
For merchants, payment failures can mean lost sales, abandoned checkouts and additional customer-service costs. An AI model tuned to payment patterns could help payment providers understand why transactions fail and improve the decisioning behind payment flows.
The wider industry question is how such systems balance higher approval rates with fraud controls. Better payment intelligence must preserve security, explainability and customer protection rather than simply approve more transactions.
BFSI NXT perspective: Payment technology is moving from processing transactions to continuously optimising the experience around them.
Related reading
- India’s Digital Payments Need the Next Layer: Reliability, Intelligence and Recovery
- DBS Bank India and Juspay Expand DBS MAX for Multi-Gateway Collections
