Google Launches Universal Commerce Protocol — Writing the Rules for AI Shopping

Google just made its biggest commerce play in years—and it’s not about ads. The company has launched the Universal Commerce Protocol (UCP), an open standard that defines how AI agents operate across retail ecosystems. Co-developed with Shopify, Etsy, Wayfair, Target, and Walmart, with endorsements from Visa, Mastercard, Stripe, and American Express, UCP positions Google as infrastructure — as explored in the economics of AI compute infrastructure — layer rather than mere discovery surface.

The Protocol Stack

UCP doesn’t stand alone. It integrates with three other protocols Google has been developing:

Agent-to-Agent Protocol (A2A): Enables AI agents from different companies to communicate and transact with each other—a retailer’s agent negotiating with a supplier’s agent, for instance.

Agent Payments Protocol (AP2): Standardizes how AI agents handle payments, authorization, and fraud detection across payment networks.

Model Context Protocol (MCP): Defines how agents maintain context across interactions, enabling coherent multi-step transactions.

Together, these protocols create a complete commerce operating system for AI agents. Google is betting on interoperability over proprietary lock-in—a strategic choice that deserves scrutiny.

Business Agent: The Interface Layer

The most visible product is Business Agent, which enables shoppers to chat with virtual sales associates from brands like Lowe’s, Michael’s, Poshmark, and Reebok. These aren’t generic chatbots—they’re customizable brand representatives with distinct voices, product knowledge, and direct purchase capability.

For retailers, Business Agent offers something valuable: direct customer relationships mediated by their own AI, rather than dependency on Google’s search algorithms. The protocol ensures their agents can operate anywhere, not just within Google’s ecosystem.

Merchant Center Evolution

Google’s Merchant Center now supports enhanced data fields specifically designed for AI agents: product Q&A, compatible accessories, substitutes, and structured comparison data. This isn’t just product catalog optimization—it’s training data for the agentic commerce era.

The Strategic Logic

Why would Google create open standards that let commerce flow beyond its own properties? Because the alternative is worse. If Amazon or Apple — as explored in the interface layer wars reshaping consumer tech — creates the dominant agentic commerce protocol, Google’s search advertising business faces existential threat. By establishing open standards, Google ensures it remains the default infrastructure provider even as shopping behavior evolves.

The Direct Offers pilot—exclusive discounts available only in AI Mode—reveals the monetization path. Google becomes the marketplace for agent-negotiated deals, taking a cut of transactions it facilitates without controlling the full stack.

What This Means for Retail

Retailers face a strategic decision: adopt UCP and gain interoperability with Google’s ecosystem, or resist and potentially face fragmented, incompatible agent experiences. The co-development partnership with major retailers suggests Google learned from past platform plays—bring partners in early rather than imposing standards unilaterally.

Google isn’t just building AI commerce tools. It’s writing the rules for how AI agents will buy and sell—and betting the industry will play along.

Source: Google I/O, Partner Announcements