Get AI summaries of any video or article — Sign up free
Google's New Universal Commerce Protocol thumbnail

Google's New Universal Commerce Protocol

Sam Witteveen·
5 min read

Based on Sam Witteveen's video on YouTube. If you like this content, support the original creators by watching, liking and subscribing to their content.

TL;DR

UCP is a retail-focused open standard meant to connect AI agents to businesses and payment providers for agentic commerce.

Briefing

Google’s Universal Commerce Protocol (UCP) is positioned as the missing retail layer in agentic commerce: a shared standard that lets AI agents discover products, add items to a cart, and complete checkout across many retailers and payment providers. The practical payoff is straightforward—retailers can expose what they sell once, and agents operating in chat and search experiences can reliably route shoppers to a purchase flow without building a separate integration for every platform.

UCP was announced at the National Retail Federation conference by Sunda Pashai, with the emphasis squarely on retail use cases rather than agent builders. The scenario described is less about fully autonomous shopping and more about “background” agent activity: a shopper chats in Google’s AI search mode (and potentially other platforms), an agent finds relevant products, presents options in the same environment, and then enables payment without forcing the user to jump through multiple sites. Google’s pitch targets businesses already comfortable with Google’s ecosystem—retailers that run ads and use Google services—making adoption feel less like a leap and more like an extension.

The protocol’s goals build on earlier agent-era standards. MCP focused on tool access, A2A enabled message passing between agents, AP2 centered on payments, and AGUI supported dynamic user interfaces. UCP is framed as an extension that connects agents to commerce systems: it’s an open standard (not a Google-only product) designed for agent-to-business and agent-to-payment-provider communication. Product discovery is highlighted as the most valuable component, since today’s agent shopping often relies on brittle web browsing and inconsistent search results. UCP aims to make product and service listings more “agent-readable,” so businesses can be found and presented without custom work for each agent or platform.

Checkout is the second major pillar. UCP is set up to support cart actions and payment handling, though the transcript leaves open whether it always relies on AP2 or can function independently. Google also claims UCP was co-developed with major commerce players including Shopify, Etsy, Target, Walmart, and Wayfair, and it has attracted additional retail partners—an intentional signal that this is meant to be broadly interoperable. Amazon’s absence stands out, but Google’s strategy appears to lean on its distribution advantage: UCP is expected to power a new checkout feature in Google products such as AI mode in Search and the Gemini app.

Beyond transactions, Google announced “business agents,” described as a way for shoppers to chat with brands directly in search—essentially a virtual sales associate that can answer questions in a brand’s voice. That move could also shift how brands handle customer inquiries and potentially support, while Google monetizes the flow by placing commerce and offers inside its answer surfaces.

The core problem UCP targets is discoverability and integration sprawl. Without a universal standard, every platform and retailer needs custom connections, and agent-driven shopping multiplies that complexity. Even with today’s app ecosystems, it’s unclear how systems decide what to recommend—whether quality, user fit, or commercial incentives dominate. UCP’s promise is universal compatibility: expose inventory and offers once, then let the protocol carry them into agent experiences.

For developers, Google has published the protocol at ucp.dev with documentation, schema references, and a playground. The open question is adoption: whether UCP becomes a widely used building block for agent developers or mainly a retailer-facing mechanism that powers Google’s AI commerce features in AI mode and Gemini. Either way, the announcement fits a broader 2026 trend—moving from experimenting with agents to monetizing them through real commerce flows inside major platforms.

Cornell Notes

Google’s Universal Commerce Protocol (UCP) is designed to standardize how AI agents interact with retail businesses and payment providers. Its biggest promise is improved product discoverability: agents should be able to find and present items reliably without relying on fragile web browsing. UCP also supports commerce actions like adding items to a cart and handling checkout, aiming to reduce the integration burden retailers face across many platforms. Google ties UCP to its distribution advantage, expecting it to power checkout in AI mode in Search and the Gemini app. Co-development with major retailers and platforms suggests the goal is broad interoperability rather than a single-vendor solution.

What problem in agentic shopping does UCP target first—discovery, checkout, or both?

Discovery is emphasized as the main win. The transcript contrasts today’s agent shopping—often “hit-and-miss” when agents surf the web—with UCP’s goal of making products and services easier for agents to locate and present. Checkout is the second pillar: UCP is set up for cart actions and payment handling, but the transcript notes it’s still early to know exactly how payment flows relate to AP2.

How does UCP fit into Google’s broader stack of agent protocols (MCP, A2A, AP2, AGUI)?

UCP is described as a retail-focused extension of earlier agent standards. MCP enables agents to access tools; A2A provides message passing between agents; AP2 centers on payments; AGUI supports dynamic user interfaces. UCP then connects agents to commerce systems—businesses selling goods/services and payment providers/digital wallets—so agent interactions can turn into real purchases.

Why does Google’s retailer strategy matter for adoption?

The transcript argues Google can leverage retailer comfort with Google services like advertising and existing integrations. That reduces friction for retailers to adopt UCP, especially when Google plans to surface checkout inside high-traffic experiences like AI mode in Search and the Gemini app, which has added hundreds of millions of users.

What does “universal compatibility” mean in practice for retailers and platforms?

Instead of building custom integrations for every platform (and every retailer building separate connections for Google Shopping, social networks, and more), UCP aims to let businesses expose how they sell once. Then the protocol should work across multiple agent and commerce surfaces, reducing integration sprawl as AI agents multiply the number of places products could be discovered.

What uncertainty remains even with a universal protocol like UCP?

The transcript highlights a decision problem: when multiple options exist, how does an agent choose what to show? It cites uncertainty around ChatGPT app recommendations (e.g., travel apps like Booking.com) and whether ranking depends on customer fit or commercial incentives. UCP may standardize the plumbing, but recommendation logic and incentives still need to be resolved.

What additional commerce-related capability did Google announce alongside UCP?

Google announced “business agents,” framed as a new way for shoppers to chat with brands directly in search. The concept is a virtual sales associate that can answer product questions in a brand’s voice, potentially handling inquiries and later supporting customer service—turning brand communication into an agent-driven channel.

Review Questions

  1. How does UCP’s approach to product discovery differ from agent behavior that relies on browsing the open web?
  2. What integration burden does UCP aim to reduce, and why does that matter more as AI agents expand into more platforms?
  3. What open questions remain about how agents decide which products or apps to recommend, even if checkout and discovery are standardized?

Key Points

  1. 1

    UCP is a retail-focused open standard meant to connect AI agents to businesses and payment providers for agentic commerce.

  2. 2

    Product discovery is the central use case: UCP aims to make products easier for agents to find and present than ad-hoc web browsing.

  3. 3

    UCP supports commerce actions such as adding items to a cart and handling checkout, with payment integration details still emerging.

  4. 4

    Google positions UCP as interoperable across platforms so retailers can expose inventory/offers once instead of building many custom integrations.

  5. 5

    UCP is co-developed with major commerce and retail partners including Shopify, Etsy, Target, Walmart, and Wayfair.

  6. 6

    Google plans to monetize UCP through checkout and offers inside AI mode in Search and the Gemini app, leveraging Gemini’s large user base.

  7. 7

    The biggest remaining challenge is not just standardization but how agents rank and choose among competing options when multiple retailers participate.

Highlights

UCP’s core promise is better product discoverability for agents—moving beyond unreliable “surf the web” behavior toward agent-readable commerce data.
UCP is framed as the commerce layer that complements MCP (tools), A2A (agent messaging), AP2 (payments), and AGUI (dynamic UI).
Google expects UCP to power checkout inside AI mode in Search and the Gemini app, turning agent answers into immediate purchasing paths.
Even with a universal protocol, ranking and incentives may still determine what agents recommend and where money flows.

Topics

  • Universal Commerce Protocol
  • Agentic Commerce
  • Product Discovery
  • Checkout Integration
  • Business Agents

Mentioned