Summary
Google has launched Universal Cart in the United States, allowing users to collect products from multiple retailers in Search and Gemini, receive price and availability alerts, and complete checkout through Google Pay. The product links Google’s AI shopping strategy with its broader push to make agentic commerce interoperable through the Universal Commerce Protocol.
A cart that sits between search and the retailer
Google has launched Universal Cart in the United States, placing a persistent shopping cart inside Google Search and the Gemini app. Users can collect products they discover across different retailers, return to those items later, and continue monitoring their shopping progress without searching for the same products again.
The product is significant because it extends Google’s role beyond discovery. Search has traditionally directed users toward retailer websites or apps, where the actual transaction takes place. Universal Cart instead creates a Google-controlled layer for gathering products, tracking purchase conditions, and moving toward checkout. The company is not presenting the service as a new retailer. Rather, it is positioning Google as an organizing and transaction interface that sits across existing merchants.
According to the published product descriptions, Universal Cart can hold items from multiple retailers and alert users when a saved product becomes cheaper or returns to stock. Participating merchants cited in coverage include Nike, Sephora, Target, Ulta Beauty, Walmart, Wayfair, and Shopify sellers. Each merchant remains the seller of record and is responsible for order confirmations and customer support. That division preserves the retailer’s role in fulfillment, even as Google takes a larger role in the path leading to a purchase.
Payment becomes part of the AI shopping layer
Universal Cart’s checkout function is connected to Google Pay and the Universal Commerce Protocol, or UCP. Google introduced UCP as an open standard intended to help AI agents and retailers interact across different stages of commerce, including checkout. In practical terms, the protocol is meant to give software agents a structured way to communicate with merchants rather than relying only on conventional web pages and manual navigation.
Google says Universal Cart can consider relevant credit-card benefits when helping users decide how to pay and can surface applicable promotional codes. This makes payment selection part of the assisted shopping experience. The system is not merely storing a list of products; it is also helping users evaluate when and how a transaction might be completed.
That shift raises an important distinction between recommendation, authorization, and execution. An AI system may organize products, monitor price changes, or prepare a checkout, but the user’s payment authorization, the retailer’s order confirmation, and the merchant’s fulfillment obligations remain separate steps. The more those steps are brought together inside an AI interface, the more important it becomes to show users what the system is recommending, what it is authorized to do, and which party is responsible if something goes wrong.
For institutional wallet and custody operators, the broader direction is relevant even though Universal Cart currently focuses on conventional retail and Google Pay rather than digital-asset payments. Transaction interfaces may increasingly be software agents, search products, or conversational applications instead of merchant pages and standalone payment screens. In those environments, permissioning, auditability, credential protection, and human review become central design questions. Higher automation does not remove the need for controls; it makes the boundaries of those controls more consequential.
Google’s Shopping Graph as a decision layer
Google says Universal Cart operates on its Shopping Graph, which contains a large body of product information. Combined with Gemini, the company is moving toward a shopping experience in which Google does more than return links. It can help interpret a request, gather candidate products, retain a user’s shopping context, and continue monitoring the conditions attached to a potential purchase.
Universal Cart follows other Google initiatives in AI-powered commerce. Google has been developing conversational product discovery in Search’s AI Mode and has introduced a tool that uses an AI agent to contact nearby stores and check whether an item is available. Universal Cart connects those discovery and availability capabilities with a durable cart and a payment path.
For consumers, the appeal is straightforward: fewer repeated searches, less switching between retailer sites, and a single place to track items that may not be purchased immediately. For merchants, the service could provide another distribution channel for products and expose listings to users with clearer purchase intent. At the same time, merchants may become more dependent on Google’s product data, ranking systems, interface decisions, and technical requirements for completing transactions.
That dependency is one of the central issues in the development of agentic commerce. Traditional e-commerce generally gives merchants direct control over their storefronts and customer relationships. A centralized AI interface can absorb product comparison, discount discovery, payment selection, and portions of customer interaction into the platform. Merchants may gain reach, but they may also have less visibility into how users choose between products or how Google determines which offers receive attention.
Google’s statement that merchants remain sellers of record helps clarify responsibility for orders and support. It does not, however, eliminate questions about the platform’s influence over rankings, recommendations, promotions, and the path to checkout. Those questions are likely to become more important as AI systems begin to mediate more of the consumer decision process.
Why the European Union matters
Universal Cart is currently available in the United States. Google’s earlier roadmap referenced Canada and Australia after the U.S. launch, with the United Kingdom planned later. The published roadmap does not mention the European Union. That omission may reflect a combination of product readiness, commercial planning, and regulatory considerations rather than a single stated reason.
The European market presents particular sensitivity around large platforms that control multiple layers of the digital economy. Concerns may extend beyond payments and data protection to whether a platform uses its search position, operating systems, or application ecosystem to favor its own services or preferred partners. Public reporting has linked Google’s European regulatory challenges to alleged self-preferencing across areas including shopping, hotels, transport, and sports results. The company has faced penalties and requirements to take corrective action.
A product that combines product discovery, ranking, discount prompts, and checkout could attract closer scrutiny because it gives Google influence over several stages of the commercial journey. Regulators and merchants may ask whether access conditions are fair, whether competitors can connect on comparable terms, and whether Google’s own services receive advantages through the interface. They may also examine how shopping and payment data are used and whether consumers can easily understand which entity is processing or supporting each part of a transaction.
The question of an EU launch is therefore not simply one of translating the product or connecting local retailers. It may require Google to address platform neutrality, data governance, consumer protection, and the allocation of responsibility when an AI-assisted purchase crosses several businesses and services.
The next test for agentic commerce
Universal Cart illustrates a broader industry sequence: search discovers products, Gemini interprets requests and provides assistance, the cart preserves context, payment tools support checkout, and an open protocol attempts to connect agents with merchants. Amazon, OpenAI, and other companies are pursuing related AI-commerce initiatives. Competition may consequently shift from who captures search traffic to who controls the decision layer immediately before a purchase.
Convenience alone, however, does not make agentic commerce infrastructure mature. Large-scale adoption will require clear rules for authorization, refunds, disputes, fraud prevention, data minimization, and cross-merchant order management. Users need to know whether an AI is only making a suggestion, preparing a transaction, or actually executing an authorized payment. They also need a clear route for resolving problems when several retailers and service providers are involved.
Universal Cart currently represents a Google-led model of centralized shopping management and payment assistance. Its launch shows how AI can move from helping users find products to maintaining an ongoing purchase context and supporting checkout. It also highlights the unresolved balance between platform convenience, merchant control, consumer visibility, and regulatory oversight.
The strategic question is not only whether an AI agent can identify a suitable product. It is whether the agent can operate within permissions that users understand, with transactions that can be reviewed, and with responsibilities that remain clear across the platform, payment provider, retailer, and customer. As more companies place agents directly in front of real payment flows, those governance requirements will be as important as the quality of the underlying recommendations.
Source: link