Summary
India's National Payments Corporation of India (NPCI) is developing a Unified Agent Protocol (UAP) to enable AI assistants to securely make UPI payments on behalf of users. The framework will authenticate and authorize AI agents, define transaction permissions, and preserve UPI's interoperability as agentic commerce gains traction. AI agents could originate from merchant apps, payment platforms, or AI assistants like ChatGPT and Claude. Low-value, repetitive purchases (groceries, daily essentials) are expected to be the first use cases. UPI processed 22.71 billion transactions worth 28.92 trillion rupees in June 2026, with 63.5% being peer-to-merchant payments.
Key Facts
- Protocol: Unified Agent Protocol (UAP) being developed by NPCI
- Purpose: enable AI agents to make UPI transactions on behalf of users
- Authentication: agents registered, verified, and authorized across UPI ecosystem
- First use cases: low-value repetitive purchases (groceries, daily essentials)
- UPI June 2026: 22.71 billion transactions, 28.92 trillion rupees
- P2M share: 63.5% of all UPI transactions
- Similar initiatives: Visa, Google, OpenAI, Pine Labs developing agentic payment protocols
- Consumer protection: chargebacks and dispute resolution need to evolve
Why It Matters
India's UAP positions the country as a potential first-mover in creating national infrastructure for AI-led payments. With UPI already processing over 22 billion monthly transactions, adding AI agent authorization could unlock a new wave of automated commerce. The protocol's design preserves UPI's interoperability while adding a trust layer for machine-initiated transactions. This could become a template for other countries building agentic payment frameworks, especially as AI assistants become capable of making purchasing decisions independently.