The company said the protocol is designed to address a key limitation in emerging “agentic commerce” models, where AI agents can browse products, compare options and make purchasing decisions on behalf of users but are unable to complete payments because UPI transactions typically require manual authentication through a PIN or approval step.
According to Pine Labs, P3P extends UPI’s existing mandate-based payment framework, allowing users to provide authorisation in advance.
Once a mandate is approved, an AI agent can initiate transactions within predefined limits and conditions without requiring additional user intervention at the time of purchase.
The protocol combines UPI mandate infrastructure with identity verification and authorisation controls provided by Grantex, alongside HTTP 402, an open web standard for machine-readable payment requests. Pine Labs said the system includes safeguards such as spending limits, audit trails and the ability for users to modify or revoke permissions.
Digital gold savings platform Gullak has become the first live deployment of the protocol. Under the implementation, users can set conditions such as purchasing gold when prices fall below a specified level, after which an AI agent can execute the transaction automatically if the criteria are met.
Electronics retailer Vijay Sales is also conducting a proof-of-concept for the technology. The company is exploring use cases where AI agents could automatically purchase products when they reach a target price set by customers.
The launch comes as AI-powered commerce platforms seek ways to automate end-to-end purchasing journeys. While AI agents can already assist consumers with product discovery and decision-making, payment execution has remained a significant hurdle due to authentication requirements embedded within existing payment systems.
Pine Labs said P3P is currently operational on UPI and that it is working with card networks to expand the protocol to card-based transactions in the future.
First Published: Jun 12, 2026 7:49 AM IST
