Google unveils new AI shopping framework as agent-led buying moves into the mainstream

Google has taken a major step toward what it calls “agentic commerce,” a future where artificial intelligence agents shop, compare and complete purchases on behalf of users, by rolling out a new open standard and a suite of tools designed to transform how people buy online.

At the centre of the announcement is the Universal Commerce Protocol (UCP), a newly launched open framework meant to let AI agents, retailers and payment systems communicate using a common digital language. The goal is to remove the friction that currently exists when different platforms and shopping tools try to work together, enabling everything from product discovery to checkout and post-purchase support to happen seamlessly through AI.

Google said UCP has been built in collaboration with major retail and payments players, including Shopify, Etsy, Walmart, Target and Wayfair, and is backed by more than 20 companies across the commerce ecosystem, ranging from Visa and Mastercard to Flipkart and The Home Depot. The protocol is designed to work alongside existing standards such as Google’s Agent Payments Protocol and other AI communication frameworks.

Also read: Meta dismisses Instagram data leak claims after password reset scare rattles users

The company plans to start deploying UCP in the US through a new checkout experience on its AI-powered surfaces, including Search’s AI Mode and the Gemini app. Shoppers will be able to buy products directly from participating retailers while still researching them on Google, using stored payment and shipping details through Google Pay and, soon, PayPal. Retailers will continue to act as the seller of record, giving them control over fulfilment while reducing the risk of customers abandoning purchases midway.

Alongside this, Google is introducing a “Business Agent” for retailers, a branded AI chatbot that appears in Search and acts like a virtual store assistant. Shoppers can ask questions about products and get responses in the retailer’s own tone and style, with the option to move directly toward a purchase. The feature is being launched initially with select US brands such as Lowe’s, Michael’s, Reebok and Poshmark, with plans to allow deeper customisation and direct AI-driven checkout later.

To help brands remain visible in this increasingly conversational form of shopping, Google is also expanding the data retailers can submit through Merchant Center. The new attributes go beyond basic keywords to include product-related questions, compatible accessories and alternatives, allowing AI systems to surface items more accurately when users describe what they are looking for in natural language.

Also read: The AI problem in agencies: Why originality is becoming a premium skill again

On the advertising front, Google has announced a pilot called Direct Offers, which will let retailers show exclusive discounts and promotions directly within AI-powered search results. The idea is to match shoppers who are ready to buy with time-sensitive or tailored deals, using AI to decide when an offer is most likely to close a sale.

Taken together, the updates signal Google’s push to position itself at the heart of the next phase of e-commerce, one driven not just by clicks and searches, but by intelligent agents that guide, negotiate and complete purchases for users. As AI becomes more deeply embedded in everyday shopping, Google is betting that open standards and tightly integrated tools will give retailers and consumers alike a smoother path from discovery to checkout.