UCP vs ACP: What the Two Agentic Commerce Protocols Mean for an Ecommerce Team

Two protocols now sit between AI assistants and store checkouts. The Agentic Commerce Protocol (ACP), published by Stripe and OpenAI in September 2025, standardizes how ChatGPT hands a purchase to a merchant. The Universal Commerce Protocol (UCP), launched by Google at NRF in January 2026 with Shopify, Etsy, Wayfair, Target and Walmart, covers the journey from discovery to checkout in Google AI Mode and Gemini.

The question an ecommerce lead actually faces is not which one to pick. It is whether any of this is work for your team. For most brands on Shopify-class platforms the answer is no, beyond keeping product information complete and feeds eligible.

What ACP is and who runs it

Stripe introduced ACP as "the first live standard that enables programmatic commerce flows between AI agents and businesses", and in December 2025 wrapped it in an Agentic Commerce Suite that Stripe says brands such as URBN, Etsy, Coach, Kate Spade and Revolve were onboarding to. Salesforce announced support in October 2025, building an Instant Checkout integration for Agentforce Commerce merchants. The same day, Walmart announced a partnership with OpenAI to let customers shop Walmart inside ChatGPT.

What ChatGPT does with ACP has shifted since. Shopify's help documentation describes ChatGPT as a discovery channel where shoppers complete the purchase in the merchant's own checkout, with no direct checkout setting to enable. OpenAI's developer documentation still publishes an agentic checkout spec and a product feed spec for merchants that integrate directly. For a brand, ACP is something your platform or payment provider supports. It is not a build.

What UCP is and who runs it

Google describes UCP on its developers blog as an open standard developed with Shopify, Etsy, Wayfair, Target and Walmart and endorsed by more than 20 partners including Adyen, American Express, Best Buy, Mastercard, Stripe and Visa. It is compatible with Google's Agent Payments Protocol (AP2) for the payment step, and merchants can integrate through APIs, Agent2Agent (A2A) or the Model Context Protocol (MCP).

Google's merchant guide is explicit about what the retailer keeps: you remain the merchant of record, keep your customer data, and use your existing Merchant Center feeds for discovery. The surfaces are AI Mode in Google Search and Gemini. Google's UCP FAQ adds that Merchant Center remains central, and that product feeds, brand assets, return policies and business contact information must be complete and up to date there.

Shopify co-developed UCP and, in its January 2026 announcement, said merchants set up their product data once and Agentic Storefronts surfaces it across ChatGPT, Microsoft Copilot, Google AI Mode and Gemini. CNBC reported in March 2026 that Gap would be the first major fashion retailer to offer checkout inside Gemini.

UCP and ACP side by side

Scope. ACP standardizes the checkout conversation between an agent and a merchant, with payment handled through Stripe or a supporting processor. UCP spans discovery, checkout and post-purchase, and treats AP2 as the payment layer inside it.

Surfaces. ACP is the path into ChatGPT. UCP is the path into Google AI Mode and the Gemini app. A shopper does not know or care which protocol is underneath, which is why the two are complementary rather than rivals for a retailer.

Who does the integration. On both, the merchant stays the merchant of record and transactions flow through existing payment processors. In practice the commerce platform or the processor implements the protocol, and the retailer's contribution is the product data and policies the protocol reads.

What an ecommerce team actually has to do

On Shopify, nothing to build. Agentic storefronts is active by default for eligible stores and manages ChatGPT, Google AI Mode and Gemini, Microsoft Copilot and Meta from one admin screen. Eligibility is the thing to check: Shopify's Catalog requirements ask for a Starter plan or higher, a store not in private mode, a title, at least one image, a price above zero, publication with an identifiable product URL, and no unlisted or hidden-from-search status.

On other platforms, ask your platform and payment provider which protocols they support and when, then confirm your Merchant Center feed is current and approved, because that is what UCP reads for discovery. The useful question for the vendor is plain: if a shopper asks ChatGPT or Google AI Mode for our category, can the assistant read our products through you, and what do you need from our feed?

What neither protocol does is choose your product. They carry a purchase after the assistant has decided what to show. That decision is made from product pages, feeds and structured data, which is why Google's retail team says a messy or incomplete Merchant Center feed means customers will not find your products in AI Mode. Protocol support is table stakes your platform provides. Being the product that gets recommended is the work that stays with your team.

Frequently asked questions

What is the Universal Commerce Protocol?

An open standard from Google, co-developed with Shopify, Etsy, Wayfair, Target and Walmart, that lets AI agents handle discovery, checkout and post-purchase steps with a merchant. It launched at NRF in January 2026 and powers shopping actions in Google AI Mode and Gemini.

How is UCP different from Google's AP2?

AP2, the Agent Payments Protocol, covers payment authorization and consent between agents, users and payment providers. UCP is broader and uses AP2 as its payment layer, as Google's developers blog describes.

Can a retailer support both UCP and ACP?

Yes, and large retailers do. Walmart works with OpenAI on ChatGPT and co-developed UCP with Google. Shopify merchants get both through platform-level integrations without choosing.

Do I need to rebuild my ecommerce stack?

No. Both protocols are designed to work with existing commerce systems. The merchant stays the merchant of record and transactions flow through established processors.

Where does MCP fit?

MCP, the Model Context Protocol from Anthropic, is a general way to connect AI models to tools and data. UCP lists it as one of the integration paths. For most retailers it matters only when an internal team builds its own agents against commerce data.

Protocols will keep changing and your platform will keep up with them. What does not change is that an assistant can only recommend a product it can read, with a price and availability it can trust. A Paz expert finds the gaps in your product information across feeds, product pages and AI answers, you approve the fixes, and the platform publishes them and tracks whether your products show up. Talk to us.