Google's "UCP" Aims for "Payments in Conversation"—AI Agents Disrupting E-commerce Norms
The standard shopping flow of "search, compare, purchase, and after-sales support" is beginning to be completed entirely within chat-based AI conversations. The Universal Commerce Protocol (UCP) announced by Google at the NRF (National Retail Federation) conference is a move aimed at becoming the "underlying common standard" for this. The goal is simple: to have AI agents move commerce across the board using a common language, rather than connecting individually to fragmented systems for every store and payment method.
1. What is UCP: A "Common Language" for the AI Agent Era
UCP is an open standard built on the premise of shopping via AI agents, designed to operate across product discovery, purchasing, and post-purchase support. Google explains that it was developed in collaboration with Shopify, Etsy, Wayfair, Target, Walmart, and others.
The key point is that the shopping experience remains connected on the backend "regardless of which AI (or which conversational UI) it starts with." For example, a user could consult with Gemini to find a product, purchase it directly within the flow of the conversation, and handle shipping, returns, and inquiries within the same context—that is the vision.
2. "Payments in Search" Becomes Reality: Implementation in AI Mode / Gemini
Google has announced that it will gradually offer US retailers an experience where users can "checkout during conversation" for product listings that meet certain criteria within Search's AI Mode and the Gemini app. The system is designed to use Google Pay for payments and shipping information saved in Google Wallet, and they have also stated that PayPal will be supported as a payment option.
2-1. UCP is not a "sole hegemony," but a design that allows for extensions
Google explains that UCP can be used in conjunction with existing agent-related protocols—such as Agent Payments Protocol (AP2) and Agent2Agent (A2A), Model Context Protocol (MCP)—emphasizing the idea that businesses and agents can "choose and assemble" the extensions they need.
3. Making "Serendipitous Encounters" the Core of Shopping: Shopify's Goal
Shopify CEO Tobi Lütke describes the value of agent-based models as "serendipity," where you encounter products you wouldn't have searched for yourself. For example, by simply conveying your "tastes and situation" in conversation, the AI leads you to an optimal solution that would be difficult to dig up with search keywords—this, he argues, is the "next commerce."
At the same time, Shopify has also announced the integration of in-conversation checkout for Microsoft Copilot, aiming to secure sales channels on any platform, based on the premise that the "conversational UI will be the main battlefield."
4. New Monetization: Inserting Discounts at the "Very Moment" of Conversation
As a consumer-facing change, Google has introduced a mechanism where brands can offer special discounts "on the spot" while a user is receiving product recommendations in AI Mode. For example, the experience envisioned is one where, at the very moment a user asks for advice on "a modern rug that is easy to clean for a dining room with high foot traffic," the AI presents both a product that meets the criteria and a discount simultaneously.
While this allows for proposals that align with the conversational context (the granularity of needs) better than traditional search ads, from the user's perspective, the boundary between "recommendation" and "advertisement" can easily become blurred. Moving forward, the transparency of display rules and how to control the excessive use of discounts and offers will be the key to trust.
5. What Will Change: The "Rewiring" of Retail, Payments, and User Experience
As UCP spreads, the competitive edge will shift from "who has the strongest e-commerce site" to (1) the starting point of the conversation (where the consultation begins), (2) execution capabilities including inventory, shipping, and returns, and (3) offer design (proposals that fit the context). Walmart's announcement of initiatives to handle everything from cart creation to purchase within Gemini is a symbol of this trend.
On the other hand, the challenges are also clear. How to remove the anxiety of AI "buying things on its own," and how to design the division of responsibility for erroneous purchases, returns, and support. UCP is a step toward organizing the "wiring diagram," but the key to its widespread adoption lies in a secure design where the user retains control (confirmation steps, visualization, and permission management).