UCP, ACP, and AP2: Why AI Agent Payments Have Three Competing Protocols (And What It Means for Builders)

UCP, ACP, and AP2 are three separate, competing protocols that let AI agents complete purchases on a merchant's behalf — and right now, none of them talks to the others. If you're building an agent that needs to pay for something, that fragmentation is the single biggest reason integration takes longer than it should.

What each protocol actually does

All three solve the same basic problem — letting an AI agent hand off cart and payment details to a merchant in a structured, machine-readable way — but they were built by different players, launched within months of each other, and aren't interoperable:

UCP (Universal Commerce Protocol) has rolled out with major US retailers including Nike, Sephora, Target, Ulta, Walmart and Wayfair, with Canada and Australia next. UK availability is still pending.
ACP (Agentic Commerce Protocol) launched with OpenAI's Instant Checkout, but adoption has been limited — only around 12 Shopify merchants went live before OpenAI retired the feature in March 2026.
AP2 is the third entrant, adding to the same fragmented landscape.

Why this matters if you're building an agent

Each protocol requires a separate, bespoke integration. A merchant supporting UCP but not ACP is invisible to an agent built only for ACP — and vice versa. Adyen took this seriously enough to build a three-protocol hedge product across its 200+ enterprise merchant base, specifically because protocol risk was the single biggest unknown on the merchant side.

For builders, the practical effect is exactly what one payments executive described in trade press: agent commerce integration today means custom plugs for custom sockets. Every new merchant or protocol is another bespoke build, not a reusable connection.

There's a second problem underneath this: most of these protocols require merchant opt-in. The long tail of ordinary online merchants — the ones without dedicated engineering teams to adopt a new checkout protocol — remains functionally invisible to protocol-compliant agents, regardless of which protocol you pick.

The underlying issue: it's not just protocols, it's readiness

Protocol fragmentation is a symptom of a bigger gap. Merchant-side agentic readiness is also held back by machine-unreadable product data, legacy stack weight, and unclear trust and liability frameworks — meaning even a merchant that picks a protocol still has real technical work ahead before an agent can reliably transact with them.

Where this leaves builders today

If you're building an AI agent that needs to pay for things, you're currently choosing between:

Picking one protocol and accepting that your agent can only transact with merchants who support it
Building and maintaining multiple integrations to cover more merchant coverage, at real engineering cost
Waiting for consolidation that may be years away, if it happens at all

None of these are good options for a small team trying to ship.

A single integration instead of three

This is exactly the gap APS is built to close. Instead of integrating against each protocol separately, APS gives your agent one integration — a universal agentic wallet — that connects to agentic commerce-enabled merchants regardless of which protocol they've adopted, with spending controls, mandates, and user approval built in from the start.

You link your wallet to any AI agent you control in a couple of minutes, no bespoke protocol work required.

FAQ

What's the difference between UCP, ACP, and AP2? They're competing protocols for the same purpose — letting AI agents complete purchases with merchants — built by different organisations and not interoperable with each other. Merchant adoption differs by protocol and region.

Do I need to support all three protocols? Not with a universal wallet approach — that's precisely the problem it's designed to remove. Direct protocol integration does require picking and maintaining each one separately.

Will these protocols eventually merge or standardise? There's no confirmed consolidation timeline. Adyen's decision to hedge across all three suggests the industry doesn't expect this to resolve quickly.