The National Payments Corporation of India (NPCI) is finalizing the Unified Agent Protocol (UAP) to enable AI assistants to execute UPI transactions autonomously. By allowing delegated, pre-authorized payments, this framework aims to streamline automated retail commerce. Investors should track how this shift from manual to AI-driven payment mandates affects digital payment infrastructure, fraud liability frameworks, and existing transaction limits.
The National Payments Corporation of India (NPCI) is preparing to launch a new framework called the Unified Agent Protocol (UAP), which will allow AI agents to perform transactions on the Unified Payments Interface (UPI) network autonomously. Currently, UPI requires a user to manually authenticate every purchase using a PIN. The new protocol aims to remove this friction by enabling AI assistants to complete payments based on pre-defined user rules and delegated authority, effectively moving the system toward what is known as 'agentic commerce.'
The framework relies on two existing UPI features to function. It uses 'UPI Circle,' which allows a user to delegate payment authority to an AI, and 'Reserve Pay,' a mechanism that ring-fences specific funds for future debits. While Reserve Pay currently supports pre-authorized spending, it is generally capped at ₹10,000 for a 90-day period. The NPCI is expected to evaluate and likely adjust these caps to support the higher frequency of transactions typical in AI-driven commerce.
This move comes as the digital payments sector sees aggressive activity from both local and global players. Private companies have already begun exploring this space, with Pine Labs launching its 'P3P' protocol earlier this year to facilitate single-authorization AI payments. Meanwhile, global payment networks like Mastercard and Visa are reportedly developing their own capabilities to capture this emerging market, indicating a strategic shift in how digital payment infrastructure is being built to accommodate automated services.
For investors and market participants, the transition to AI-driven payments introduces new operational and security variables. A primary concern is the increased risk of fraud if 'rogue' AI agents or system errors lead to unauthorized transactions. Consequently, the success of this protocol will depend on the establishment of a robust liability framework, mandatory identity verification, and granular audit trails to track AI-led activity.
Another important monitorable is how the Reserve Bank of India and NPCI handle transaction limits. As AI agents begin to handle more complex tasks, the current constraints on pre-authorized funds may need to be expanded, which could impact the volume and size of digital transactions. Market observers will be looking for further updates on the official launch timeline, which is expected to coincide with upcoming industry events like the Global Fintech Fest, and whether existing banking infrastructure can scale to handle the potential surge in autonomous, high-frequency payment requests.