Google's Universal Cart and UCP protocol are accelerating AI-driven commerce. What it means for product visibility and export prospecting for European SMBs.
Agentic commerce, product visibility and new opportunities for European exporting companies.
On October 15, 2026, Google began rolling out Universal Cart in the United States, within Google Search and the Gemini app. This smart cart brings together products from multiple retailers and makes it easier to find your selections again, track price changes and move more easily toward checkout. Source: Google, October 15, 2026.
This might look like a simple convenience upgrade to online shopping. But it signals something deeper: discovery, comparison and part of purchase preparation are migrating toward AI assistants.
For brands and exporting companies, the question becomes strategic. It is no longer enough to attract visitors to your site: you also need to let third-party systems discover your products, understand their value, and verify that they can actually be bought.
1. Universal Cart: a cross-retailer cart rather than another storefront
In traditional e-commerce, a shopper moves from a search to several different stores, building a separate cart on each site.
Universal Cart changes that journey. A user can keep items spotted at different merchants inside their Google cart and find that selection again across several Google interfaces. The system can also flag price drops, restocks and certain promotions.
Google presents Universal Cart as an experience tied to its Universal Commerce Protocol (UCP), an infrastructure meant to make exchanges easier between platforms, merchants and payment providers.
Two things need to be kept separate, though: Universal Cart is a consumer-facing interface; UCP is a technical protocol. One being rolled out doesn't mean every merchant immediately has access to everything the other makes transactionally possible.
The essential shift is elsewhere: the consumer can now keep part of their selection journey inside an environment that no single brand controls.
2. The real break: AI becomes a decision intermediary
Today, a search engine mainly helps find pages. A shopping assistant can go further: interpret a need, compare several offers and propose a reasoned selection.
Picture a German buyer phrasing this request:
Find me three French manufacturers of durable professional furniture that can deliver to Germany, with reliable spec sheets and a budget under âŹ5,000.
The assistant doesn't just need a well-ranked page. It has to identify comparable products and access useful information: materials, dimensions, certifications, price, availability, countries served or lead times.
Visibility is therefore no longer just a position in search results. It also depends on an offer's ability to be interpreted, compared and retained by AI systems.
This evolution extends SEO and GEO (Generative Engine Optimization), but adds an operational dimension: a recommended product must be orderable under credible conditions.
3. UCP: a common language for agentic commerce
Unveiled by Google on January 11, 2026, the Universal Commerce Protocol is an open standard designed to connect AI-driven shopping experiences to merchants' systems. It was developed with industry players, including Shopify, Etsy, Wayfair, Target and Walmart.
Its ambition is to reduce the proliferation of one-off connections between assistants, catalogs and payment services. The protocol is designed to cover several stages of the commercial journey, from discovery to after-sales, depending on the capabilities each participant actually integrates.
For a company, the lesson is simple: its e-commerce system must progressively be able to communicate reliable commercial information to external interfaces.
This notably requires:
- identifiable product references and variants;
- prices and stock levels consistent with the merchant site;
- usable shipping and return information;
- understandable technical attributes;
- interfaces or feeds compatible with the channels actually used.
The point isn't to build a custom UCP integration right away. It's first to have clean, interoperable commercial data.
Source: Google Developers, UCP technical overview.
4. Google and ChatGPT: not identical shopping journeys
The rise of agentic commerce doesn't mean every platform is adopting the same model.
Google is working to connect its search engine, Gemini, its shopping infrastructure and UCP's capabilities. Universal Cart lets it stay with the user longer during product selection.
ChatGPT, for its part, is developing product discovery and comparison based on public information and feeds supplied by merchants. In its current communication aimed at merchants, OpenAI favors product discovery and purchase completion on merchants' own sites or apps, rather than a generalized autonomous checkout flow inside ChatGPT.
This difference matters. Companies shouldn't assume the same flows, rules or features will apply everywhere.
Source: OpenAI, merchant information.
5. Export: do you still need to translate your entire catalog?
Not necessarily â and that may be one of the more interesting consequences of this shift for exporting SMBs.
A multilingual AI agent can understand a request made in German and analyze product information written in French or English. That reduces, at least partly, the dependency between the language used to discover a supplier and the language of its original catalog.
A French company could therefore be identified by a foreign buyer even without a fully translated version of every product sheet in each target market's language.
That said, understanding a product isn't the same as being able to display or sell it on every platform.
Google Merchant Center lets you target certain markets with a feed in a supported language, even if that language isn't the country's primary one. It still requires consistency between the language of the product data, the landing pages, and the applicable checkout flow. Each channel's own rules still matter.
Three levels need to be distinguished:
- AI discovery: a catalog can be understood beyond its original language; an exhaustive translation isn't always a technical prerequisite.
- Distribution on platforms: each service applies its own requirements for language, format and data consistency.
- Conversion and customer relations: essential information, contractual terms and after-sales service must remain understandable to the buyer and compliant with the target market's obligations.
The right strategy, then, isn't necessarily to mechanically translate thousands of sheets. It's rather to structure data at the source, then localize first the content and steps that actually influence the sale.
Source: Google Merchant Center, supported languages and currencies.
6. For SMBs, a new way to approach international markets
Take the case of a French professional lighting manufacturer wanting to sell in Italy and Germany.
In a traditional international strategy, it might plan localized sites, SEO and advertising campaigns, and separate prospecting efforts for each country.
Those investments can still be worthwhile. But AI assistants introduce a complementary channel: a potential customer can phrase a need directly in their own language and ask the assistant to identify suppliers able to meet it.
To be selected, the manufacturer mainly needs to provide the elements that let its offer be qualified: technical compliance, dimensions, materials, lead times, warranties, availability and shipping terms abroad.
AI could thus lower some of the costs of international discovery, without removing the logistical, commercial or regulatory obligations specific to export.
The opportunity is particularly interesting for companies with differentiated products but limited marketing budgets.
7. The flip side: who controls the customer relationship?
When the consumer starts their journey inside an AI assistant, the merchant can lose part of their visibility into the questions asked, the products compared, or the criteria that drove the selection.
This shift raises several challenges:
- Marketing attribution: identifying the recommendations and interactions that actually led to a sale.
- Dependency: avoiding concentrating acquisition on a single assistant or platform.
- Differentiation: not letting a product be reduced to its price and a handful of technical attributes.
- Data quality: avoiding recommendations based on outdated prices or incorrect stock.
- Customer relationship: preserving a quality brand experience and after-sales service, even when the purchase journey starts elsewhere.
It would be an exaggeration to say assistants have already replaced online stores. Still, they are becoming a new arena where offers compete â and potentially a new point of passage between brands and their customers.
8. The European market: watch the rollout without waiting to prepare
As of October 22, 2026, Universal Cart is deployed in the United States. Equivalent availability in the European Union shouldn't be assumed without a specific announcement from Google.
The European context adds important dimensions: consumer protection rules, personal data, digital competition, pre-contractual information and requirements specific to the products sold.
For a European company, however, preparation shouldn't depend solely on the launch date of an American service.
A structured, reliable catalog remains useful for SEO, marketplaces, Shopping campaigns, comparison sites, distributors and future AI agents.
This is therefore a cross-cutting investment, not a bet placed exclusively on Universal Cart.
9. Five concrete priorities to prepare your business
1. Audit your product data
Check the completeness of attributes, price consistency, stock freshness and variant identification. Errors that are tolerable in a marketing sheet become problematic once machines are automatically comparing offers.
2. Make your information accessible and structured
Use appropriate structured data, documented merchant feeds and, where relevant, secure consultation interfaces. Precision matters more than multiplying pages with little differentiation.
3. Test international discoverability
Submit realistic purchase requests to different assistants in English, German, Spanish or Italian. Observe whether the company appears, which products are selected, and what information is missing. These tests provide signals, but don't guarantee a stable ranking.
4. Localize selectively
Don't automatically translate the entire catalog without measuring the commercial upside. Start with strategic products, high-intent pages, regulatory information, shipping terms and conversion steps.
5. Prepare for cross-border transactions
Clarify countries served, currencies, applicable taxation, shipping costs, returns and customer service. An international recommendation only creates value if the customer can actually buy.
Conclusion: from visibility on Google to the ability to be chosen by an AI
Universal Cart illustrates a gradual shift in online commerce: the search engine no longer just sends a visitor to a site; the assistant helps build the purchase decision.
For European companies, especially exporting SMBs, this transformation can open new prospecting and sales channels. The ability of AI models to work across languages can reduce certain barriers to international discovery.
But the decisive factor remains the availability of accurate, structured and commercially usable product data.
The challenge, then, isn't only to be found. It's to be understood, recommended and ready to sell, whatever country the buyer is in or interface they use.
Busony: preparing businesses for agentic commerce
At Busony, we help companies adapt their visibility and sales processes to the arrival of AI agents: answer-engine optimization, information structuring, agentic integrations and international development.
Is your catalog ready to be understood and recommended by AI assistants?
Discover our approach at busony.com.
Sources
- Google, "An intelligent new shopping cart," October 15, 2026: https://blog.google/products-and-platforms/products/shopping/universal-cart-launch/
- Google Developers, "Under the Hood: Universal Commerce Protocol," January 11, 2026: https://developers.googleblog.com/under-the-hood-universal-commerce-protocol-ucp/
- Google Merchant Center, "Supported languages and currencies": https://support.google.com/merchants/answer/160637?hl=en
- Google Merchant Center, "About landing page requirements": https://support.google.com/merchants/answer/4752265?hl=en
- OpenAI, "Power product discovery in ChatGPT": https://chatgpt.com/merchants/
- OpenAI Developers, "Product feeds": https://developers.openai.com/commerce/specs
Editorial note: the international purchase scenarios presented in this article illustrate possible uses. They do not represent universally deployed features, nor a guarantee of recommendation by AI assistants.