NPCI reportedly readies UPI protocol for AI-led, delegated payments

NPCI is reportedly set to launch an agentic-payments protocol on UPI, allowing AI agents to make capped purchases such as grocery orders. The initial focus is expected to be low-value, high-frequency transactions, with merchant integration infrastructure and potential AutoPay mandate portability to follow.

— Source publishedTue, 1 Sept, 2026, 16:09 IST·First seen Tue, 1 Sept, 2026, 16:24 IST·Source Inc42 · Buzz

What happened

NPCI is reportedly set to launch a UPI agentic-payments protocol enabling AI agents to make delegated, capped purchases such as grocery orders. The framework

Key facts

  • 2026
  • 2025

Why this matters

Evaluate partnerships or acquisitions in UPI orchestration, merchant integration, consent management, and AutoPay portability to secure a position in India’s emerging delegated-commerce stack.

What to watch

  • Formal NPCI product announcement, operating guidelines and launch-partner list.
  • Rules on consent, transaction caps, recurring permissions, revocation, authentication and liability for erroneous agent purchases.
  • Whether UPI AutoPay mandates can be ported across merchants, apps or AI agents.
  • Bank support for delegated-payment credentials and real-time customer notifications.
  • Merchant API standards for catalogue retrieval, inventory checks, substitutions, refunds and order-status updates.
  • Early integrations by quick-commerce, grocery, food-delivery, marketplace, bank and conversational-AI platforms.
  • Evidence that agent-originated orders raise repeat frequency or basket value without increasing cancellation, return or support costs.
  • Regulatory or consumer-protection action following disputed delegated transactions.
  • Prioritize machine-readable product, inventory, price, promotion and substitution data for high-frequency replenishment categories.
  • Define delegated-order guardrails: SKU eligibility, spend caps, quantity limits, delivery windows, price-change tolerances and escalation rules.
  • Test an agent-ready repeat-order flow for loyalty members, starting with staples and frequently repurchased baskets.
  • Review payment, refund, cancellation and customer-service workflows for orders initiated by an agent rather than directly by a shopper.
  • Negotiate with payment providers and commerce platforms for access to agent-payment telemetry, mandate portability and attribution data.
  • Protect margin by setting promotion rules that prevent agents from automatically stacking discounts, coupons and loyalty redemptions beyond approved thresholds.

Also reported by