Gnani.ai, Razorpay launch agentic AI platform for in-call UPI payments

The new platform allows Gnani's AI agents to initiate and complete UPI payments during live customer calls, using Razorpay's MCP server to connect directly to its payments infrastructure

Indian AI startup, Gnani.ai, has partnered with payments platform Razorpay to launch an agentic AI collections platform that can complete payment transactions during live customer calls, the two companies said on Wednesday. The new platform combines Gnani.ai’s agentic AI systems with Razorpay’s Model Context Protocol (MCP) server, which allows an AI agent to initiate and finish a payment within the same conversation. 

"For the first time, enterprises can automate collections at scale without sacrificing conversion," said Ganesh Gopalan, Founder and CEO of Gnani.ai. "This integration delivers what voice AI has promised but never achieved: the ability to complete transactions, not just initiate conversations. By leveraging Razorpay's Model Context Protocol, we've transformed voice AI from a reminder system into an action-completion platform that delivers both automation economics and human-like effectiveness."

How the platform works

According to Gnani AI, the system connects Gnani’s agentic AI Systems, which the company said already handles more than 10 million calls a day, with Razorpay’s MCP server, an AI-native layer that connects AI agents directly to Razorpay’s payments infrastructure. The company said that this gives the AI a real-time interface to initiate and track payments. During a call, the AI agent can assess a customer’s intent, choose an appropriate payment method, generate a payment link or a UPI collect request, and monitor the transaction status while the conversation is still ongoing. 

The platform currently supports two payment flows: one-time UPI payments and mandate-based UPI SBMD (also known as Reserve Pay) payments for recurring obligations. The companies said support for other instruments such as cards and wallets is planned.   

Khilan Haria, chief product officer at Razorpay, said, “Agentic AI only becomes truly valuable when it can act, not just converse. With Razorpay’s Agentic Payments infrastructure powering this integration with Gnani.ai, we’re giving AI systems a secure, real-time pathway to complete payments inside live customer conversations. This moves collections from ‘follow-up’ to ‘finished’ in a single interaction, unlocking a completely new operating model for financial institutions where intent, authorization, and settlement happen seamlessly and safely in one flow.” 

Gnani.ai said its platform handles the full workflow, including verifying outstanding amounts, confirming payment intent, generating secure transaction-specific payment links, sending them over channels such as SMS or WhatsApp, tracking payment status and confirming receipt during the same call. The company said its systems support conversations in more than 38 languages. 

On security, the companies said customer payment details do not pass through the voice AI system, and payments are processed through Razorpay’s infrastructure. They also said payment links are generated with transaction-specific parameters to prevent reuse, and mandate-based payments follow existing UPI security protocols. 

Part of a wider push on agentic payments

The Gnani.ai–Razorpay partnership builds on Razorpay’s earlier work with the National Payments Corporation of India (NPCI) and OpenAI to bring agentic payments to ChatGPT. That pilot, which is currently in beta, allows users to complete UPI transactions inside a chat interface using UPI Reserve Pay, and also relies on Razorpay’s MCP server to connect the AI system to its payments stack. 

At the India AI Impact Summit 2026, NPCI’s executive director for growth, Sohini Rajola, told Business Standard that the organisation plans to work with more AI platforms to expand UPI payments through AI agents.\

Availability

Gnani.ai and Razorpay said the agentic AI collections platform is available immediately for enterprise deployment. The companies said financial institutions and enterprises can access the system through Gnani.ai and schedule technical demonstrations as part of the rollout.