India Building AI Agent Registry for Digital Payment Network

India's National Payments Corporation is developing a registry system to authenticate and oversee artificial intelligence agents that conduct financial transactions on the country's digital payment infrastructure, according to three sources familiar with the initiative.

The registry will serve as a core component of NPCI's planned Unified Agentic Protocol, vetting AI agents authorized to execute payments on behalf of users. Reuters first reported the development, citing sources who could not be named because they lack authorization to speak publicly.

Why it matters

As AI agents gain the ability to autonomously handle financial transactions, payment networks face new security and accountability challenges. India's registry approach offers a potential model for balancing innovation with oversight—a question payment systems worldwide are now confronting as agentic commerce emerges.

Registry scope and expansion plans

The verification system will initially focus on AI agents operating within the Unified Payments Interface, India's dominant retail payment network. A senior industry executive told Reuters the registry could later extend to credit and debit cards, bill payment systems, and other transaction methods.

NPCI operates UPI and other retail payment systems across India. The organization did not respond to requests for comment on the registry plans.

Starting with small transactions

The agentic payment capability on UPI will first enable AI agents to process small, recurring digital payments without requiring user approval for each transaction. Groceries and other low-value purchases are expected among the earliest use cases.

Over time, the system could handle more sophisticated instructions, according to a second source. Examples include purchasing products when discounts reach specified thresholds or executing investment trades when prices cross predetermined levels.

Regulatory questions remain

While the registry will index authorized agents, critical questions around liability for erroneous or unauthorized payments still require regulatory frameworks, a third source noted. These governance issues parallel concerns that have surfaced at major AI companies.

The registry development comes amid increased scrutiny of problematic agent behavior at firms including OpenAI, Meta, and Anthropic.

Global context

India's initiative reflects broader international efforts to establish identification and authentication systems for AI agents in commerce. China's Payment & Clearing Association, supervised by the central bank, has issued guidelines covering agent identification, security protocols, and risk controls.

Major technology companies including Amazon, Google, and Microsoft have also built agent registries to support agentic commerce on their platforms.

Reuters reporters Ashwin Manikandam and Jaspreet Kalra first reported these details.