India's Agentic Payments Infrastructure: P3P, UPI ReservePay, OTM, and What It Means for Merchants

India just quietly built the most sophisticated agentic payments infrastructure in any major market. Pine Labs' P3P (Pine Labs Payment Protocol) running on NPCI's UPI ReservePay and OTM rails, within RBI's e-mandate framework, with an identity + audit layer called Grantex, gives Indian merchants agent-completable transactions with a level of programmable authorization that the ACP/UCP/AP2 stack in the US and Europe is still catching up to. Here's what it is, how it fits the Layer-03 payments stack in the 7-Layer Map, and what merchants building in India should do about it.

What P3P is, in one paragraph

P3P (Pine Labs Payment Protocol) is a merchant-facing payment protocol that enables an AI agent to execute a transaction on a consumer's behalf, within pre-set limits, at a moment the consumer isn't present. The consumer approves a scoped mandate once — a spending cap, a category constraint, a specific condition trigger — and the agent executes the transaction when the trigger fires. P3P runs on NPCI's UPI ReservePay (for reservation-style holds) and OTM (One-Time Mandate) rails, within the RBI's e-mandate framework. Every P3P transaction carries a Grantex-verified identity + scoped authorization + immutable audit trail.

That's the technical structure. The strategic implication is bigger: India has built the first widely-deployed agentic payments infrastructure in a major market. What ACP + AP2 + TAP are still architecting in the US-and-Europe ecosystem, UPI + P3P + Grantex is already executing in India in production.

Where P3P sits in the agentic commerce stack

Using our 7-Layer Map framework:

P3P is meaningfully novel because it combines Layer 03 (payments authorization) with Layer 07 (trust verification) into a single mandate primitive. The consumer's UPI approval creates a scoped, time-limited, revocable authorization that both authenticates the agent and defines what the agent can do. Grantex enforces the scope; UPI ReservePay/OTM executes the payment.

That combination is what's structurally distinctive about the India approach. In the US ACP model, the agent identity (TAP), payment authorization (AP2), and payment execution (Stripe delegated tokens) are three separate integrations that a merchant has to reason about independently. In the P3P + Grantex + UPI model, they collapse into one primitive that a Shopify-comparable merchant integration can access as a single mandate.

The rails underneath — UPI ReservePay and OTM

Two NPCI rails do the work.

UPI ReservePay — a reservation-style mandate where funds are held in the consumer's account (not debited) until an execution condition is met. Analogous to a card authorization hold. Enables use cases where the agent needs to commit to a purchase (e.g., reserve an iPhone at launch) but the final debit happens when the merchant confirms shipment.

UPI OTM (One-Time Mandate) — a pre-authorized single-execution debit that fires when a condition is met. Analogous to a stored card charge, but scoped to a specific merchant, amount, and time window. Enables use cases where the agent needs to execute a transaction the moment a condition triggers (e.g., buy gold when price drops below threshold).

Both rails sit within RBI's broader e-mandate framework — the regulatory infrastructure that governs recurring and conditional payments in India. The e-mandate framework is what makes agent-initiated transactions legally executable in the first place. Grantex ensures each P3P transaction complies with the mandate the consumer authorized.

Live use cases (as of H1 2026)

Four are in production, according to public Pine Labs disclosure:

Gullak — gold savings on price triggers. Consumer sets a target price per gram (e.g., ₹7,200) and a monthly cap (e.g., ₹2,000). Agent monitors gold prices; when the trigger fires, executes ₹500 purchase within cap. Consumer wakes up to notification + gold in account.

Vijay Sales — flash-sale mandates for consumer electronics. Consumer approves an OTM mandate for a specific product category (e.g., "reserve iPhone launch, block up to ₹20,000 down payment"). Agent monitors the merchant's availability trigger. The instant the product releases, the agent blocks the down payment via UPI OTM, securing the order.

Cult.fit — subscription re-activation on qualifying offers. Consumer approves a re-activation mandate during their active subscription period. After they churn, the agent monitors for qualifying re-activation offers. When one appears within the pre-approved budget, the agent re-subscribes and the consumer receives a notification.

Emerging: SEBI-broker integrations — systematic investment strategies executed by agents ("buy 5 shares of Infosys every time it drops below ₹1,400 up to ₹30,000/month"). The broker remains the principal for every trade under SEBI's algo trading framework; P3P handles the payment execution layer.

Additional use case shapes surfaced in the public discussion but not yet in production include NBFC EMI optimization (Account Aggregator balance signal + P3P debit trigger for early-stage NPA prevention), usage-based AI startup billing, and travel fare monitoring with agent execution.

Why India got here first

Three structural conditions made India the first-mover.

UPI's account-mandate model. UPI supports account-level mandate primitives (ReservePay + OTM) that most other national real-time payment systems don't. Brazil's PIX, EU's SEPA Instant, and the US's FedNow all support real-time settlement but don't (yet) have the mandate + scope + audit layer P3P builds on.

Regulatory clarity via RBI's e-mandate framework. Agent-initiated payments live in a regulatory gray zone in most jurisdictions. RBI's e-mandate framework provides clear rules for what a consumer can pre-authorize, how mandates can be scoped, what audit requirements apply. That regulatory clarity unlocks merchant confidence to build on the primitive.

Pine Labs' incumbent PG position. Pine Labs already processes a meaningful fraction of Indian merchant transactions as a Payment Aggregator. That distribution + relationships + engineering capacity made them the natural implementer of the merchant-facing protocol layer. The Grantex identity + audit infrastructure is what makes the primitive trustable at scale.

Comparable non-India infrastructure

Razorpay and Cashfree are both building comparable capabilities in India. Razorpay's agent-payment tokenization work + Cashfree's mandate authorization APIs are structurally similar to P3P — different implementations of the same primitive. Expect competitive protocol positioning in India through H2 2026 and standardization pressure from RBI + NPCI shortly thereafter.

Globally: the closest architectural analog is the combination of ACP (OpenAI/Stripe) at the protocol layer + AP2 (payments authorization) + TAP (trust verification). Stripe's delegated payment tokens for ACP are the closest US primitive; they're structurally comparable but not yet as unified as P3P + Grantex + UPI. Basis Theory, Nekuda, and Skyfire all build adjacent capabilities.

The strategic implication: India's agentic payments infrastructure is currently ~12-18 months ahead of the US on the authorization primitive. That gap will close as US-and-Europe stack matures — but for merchants operating in India, the window to be an early adopter is open now.

What merchants should do

Five use case shapes that map well to P3P today.

1. Flash-sale or scarcity-driven categories (electronics launches, limited drops, concert tickets, high-demand travel). The Vijay Sales pattern generalizes. Consumers commit ahead of time; the agent executes at the scarcity moment.

2. Price-triggered investment or savings. Gold, equities, mutual funds, SIP-adjacent behaviors. The Gullak pattern generalizes to any pre-approved investment thesis with a specific price condition.

3. Subscription reactivation. The Cult.fit pattern is repeatable across any subscription business with a churned-but-not-lost customer base. The agent monitors for qualifying offers and executes reactivation within pre-approved parameters.

4. EMI optimization (for NBFCs). Combine Account Aggregator balance signal + P3P mandate to execute EMI debits at the optimal moment for both the consumer's cash position and the lender's NPA metrics.

5. Usage-based billing for SaaS + AI + developer platforms operating in India. Consumer approves monthly cap; platform bills against actual consumption at threshold triggers. No monthly invoicing; no chasing receivables.

The strategic significance

Every payment system before P3P assumed the consumer is present at the moment of transaction. Enter PIN. Tap confirm. Complete checkout. Human must be there.

P3P + Grantex + UPI ReservePay/OTM is the first widely-deployed infrastructure that assumes the consumer isn't present at the moment of transaction — while preserving the properties (identity verification, scoped authorization, audit trail, revocability) that make the absent-consumer model actually trustable.

That shift matters for two reasons.

First, it unlocks demand that currently doesn't convert. The consumer who intended to buy the iPhone on launch but was in a meeting. The investor who intended to buy gold at ₹7,200 but was asleep. The traveler who was going to book Mumbai-Bangalore at ₹8,400 but missed the window. All of that demand exists; the infrastructure to capture it did not. P3P captures it.

Second, it changes what "agentic commerce" means practically. The US-and-Europe stack today handles agents that shop when the consumer asks. P3P handles agents that shop when the pre-approved condition triggers — with the consumer having given permission ahead of time. That's a structurally more powerful primitive, and the ranking model for which brand wins the "agent's default choice" slot becomes even more strategically valuable when the agent is executing autonomously.

Where this fits into Tru Commerce's stack

For our Indian customers and for global brands operating in India, P3P integration is a Layer-03 (payments) capability we'll increasingly expose alongside our ACP/UCP/AP2 abstraction. The merchant's own stack doesn't need to reason about which protocol layer to route through — the abstraction handles it. A consumer in California uses ACP + Stripe delegated tokens; a consumer in Bengaluru uses P3P + UPI OTM; the merchant experiences one integration.

That's the load-bearing case for a translation layer: as more agentic payments primitives emerge geography-by-geography (P3P in India, ACP + AP2 in US, comparable emerging primitives in Brazil + EU + APAC), merchants shouldn't be managing them individually. One abstraction, multiple protocols, one merchant integration.

CTA

If you're operating in India and have any of the five use case shapes above, P3P integration is worth exploring in H2 2026. To understand where your brand sits on visibility across the AI surfaces that will increasingly drive P3P-triggered transactions, start with a free Citation Rank scan.

If you're a global brand thinking about India-market strategy or a payments infrastructure company evaluating the emerging protocol landscape, book a demo — we're actively tracking the India stack maturation and can share the operational implications for global merchants.

Editorial note: this analysis draws on public disclosure from Pine Labs (Tanya Naik, Head of Online & Omnichannel, June 2026) about P3P and Grantex, plus our own field observations from Indian merchants and payment aggregators. Pine Labs' work on P3P + Grantex is a category-defining move — this is our synthesis of what it means for merchants and how it fits the broader agentic commerce stack.

— The Tru Commerce team (formerly Asva AI)