Razorpay Launches AI Payments Model to Make Digital Transactions Smarter and Safer

Fintech company Razorpay has launched an AI-powered payments foundation model designed to make digital transactions more reliable, secure and efficient as India’s digital economy continues to grow.

Named Razorpay Vulcan, the model has been developed using technology from NVIDIA and Amazon Web Services (AWS). Razorpay described it as India’s first transformer-based AI foundation model built specifically for the payments ecosystem.

The company said Vulcan is designed to address some of the biggest challenges faced by digital payments, including transaction failures, fraud, inefficient payment routing and friction during the checkout process. The technology is expected to help businesses make faster and more informed decisions about how individual transactions should be processed.

Razorpay said the development of the model was supported by an internal study involving around 1.5 million shoppers and more than 51,000 businesses. The study found that payment-related problems remain a concern not only for consumers in major cities but also for customers in smaller markets.

The company has already tested components of the model across around 3 trillion data points from its payments network. These tests have focused on areas such as transaction routing, fraud detection and risk assessment using live payment scenarios.

According to Razorpay, the technology has delivered promising results during these deployments. Payment success rates have improved by up to 10 per cent, while the system detected and prevented eight times more international card fraud. It also identified five times more potentially fraudulent or disputed transactions without increasing the number of alerts generated for businesses.

The company said the AI system has also improved the checkout experience. On Razorpay's Magic Checkout platform, 40 per cent more shoppers were able to see their preferred UPI application, helping facilitate an additional 1-2 lakh purchases every month.

Razorpay CEO and Co-founder Harshil Mathur said the broader objective was to make digital payments more dependable, particularly for consumers who are still weighing digital transactions against the familiarity of cash.

The technology could become increasingly important as India's payments landscape becomes more complex. Consumers can now pay through UPI, cards, net banking and wallets, while transactions may pass through hundreds of banks, payment gateways and other intermediaries.

Razorpay said Vulcan is designed to analyse these options in real time and identify the most suitable payment route before a transaction is attempted. This could help reduce failed payments while also improving the overall experience for customers and merchants.

NVIDIA's Pahal Patangia, Head of Global Industry Business Development and Payments, said India's rapidly evolving digital economy presents an opportunity to make payment systems more intelligent, reliable and secure.

AWS India and South Asia's Head of FSI and Conglomerates, Kiran Jagannath, said the model brings together billions of transaction insights into a unified intelligence layer. Built using Amazon SageMaker, the system is intended to reduce dependence on multiple fragmented machine-learning models and allow payment intelligence to evolve continuously.

Razorpay said the model will eventually be expanded beyond routing and fraud detection to support other critical payment decisions, including authentication, risk management and lending.

The launch marks another step in the growing use of artificial intelligence within India's financial technology sector. As digital payments become increasingly central to everyday commerce, AI-based systems that can reduce failures, detect fraud and personalise transaction decisions could play a significant role in improving both consumer confidence and business efficiency.