Zalando wants its customers to buy products from other retailers without leaving the Zalando app, and an AI shopping agent is intended to complete the purchase for them. In its 3 November 2026 announcement, the European fashion platform introduced Zalando Extended, with home-furnishings retailer Westwing as its first pilot partner. The AI checkout is planned to operate from 2027, not a feature that every Zalando customer can use today.
It is an intriguing reversal of familiar online shopping. A retailer normally tries to sell what appears in its own catalogue. Zalando is proposing to show items it does not itself sell and let software make the cross-shop transaction. To the shopper, the attraction is obvious: find a coat, then a chair or lamp, without opening another account or restarting a search. Behind the screen, somebody still has to decide who sells the item, who takes payment and who deals with a problem.
Zalando Extended could put several retailers behind one shopfront
Zalando already uses AI to help customers search its existing fashion and lifestyle offering. The new proposal goes beyond conversational product discovery. It brings items from outside shops into the app, with Zalando’s agent acting on the customer’s behalf when a purchase is made. Westwing’s home and living range provides the first example, and Zalando says it intends to broaden the model over time.
The company has a commercial reason to experiment. It reported 63.2 million active customers in the third quarter and says more than 22 million unique users had visited its Zalando Assistant during 2026. Additional categories and partner retailers could encourage shoppers to spend more time inside one service. Participating shops, meanwhile, could gain customers without building a separate technical connection for every new shopping tool. Zalando also sees scope for advertising products through its retail media business.
None of this establishes that the checkout will be frictionless in practice. A customer ordering a large item from one shop and a small item from another may have different delivery dates, fees, return rules and customer-service contacts. The convenience of one screen can make those differences less visible. What looks like a simple click will still be a collection of separate commercial obligations that must be clearly presented.
The difficult bit begins after the recommendation
Most AI shopping systems are relatively safe when they suggest alternatives. A wrong recommendation wastes time; an incorrectly placed order can waste money and create a dispute. LiveAIWire has already explored why shoppers are more comfortable with AI comparisons than AI checkout. Zalando’s proposal arrives at precisely that point of hesitation: making a piece of software trustworthy enough to act rather than merely advise.
The European Commission’s guidance on buying online sets out information and consumer rights that remain relevant however clever a shopping assistant becomes. Buyers need to know the identity of the seller, the full cost, applicable delivery conditions and the steps available if goods are faulty or unwanted. An AI agent does not make these responsibilities disappear. The precise allocation of duties in Zalando Extended will depend on its final contractual and operational design, which the announcement does not fully establish.
A sensible consumer interface might therefore retain a clear last review before ordering. The shopper should be able to see the merchant, final price, quantity, delivery address and return terms, then correct anything the assistant misunderstood. Removing pointless form-filling is attractive; removing the chance to spot a mistake is less obviously helpful. The distinction is especially sharp for bulky home purchases where a delivery error can be difficult to undo.
More choice also means more influence over what gets shown
The other unresolved issue is how the AI decides which product deserves attention. A conversation that feels like impartial advice can also be a route to advertising or profitable partner sales. Zalando says participating retailers will be able to promote items through its advertising business. That is not in itself evidence of unfair recommendations, but it raises a transparency question shoppers already face in traditional search results.
The publication has previously examined what UK watchdogs say about shopping agents and existing rules. The principle is that novel software should not obscure who is responsible for a commercial decision. The buyer still needs enough information to judge the offer and recognise when a recommendation is sponsored.
Zalando Extended is therefore more than an extra catalogue. It is a test of whether a familiar shopping app can take responsibility for a customer’s journey across other sellers while keeping the transaction intelligible. The company has announced the direction and its first partner, but the public evidence does not yet show how customers will experience disputes, corrections or returns.
There is a final commercial question for Zalando itself. The more buying decisions take place inside one interface, the more valuable that interface becomes to retailers seeking attention. Smaller shops might gain a route to customers they could not otherwise reach, yet they might also become dependent on how an intermediary ranks and displays their products. That balance will be worth examining once the pilot has real transaction data, rather than relying solely on the announcement of a planned service.
Watch the checkout, not the demonstration
When the system launches, the meaningful questions will be practical. Can customers choose whether the agent pays? Can they inspect and amend an order? Is a Westwing purchase clearly identified as a Westwing purchase? Will the helpdesk know which shop owns the problem? These details will determine whether the idea feels like a useful extension of online retail or an additional layer between the shopper and the seller.
The attraction of AI commerce is not that a machine can press the buy button. Ordinary software has done that for years. The challenge is letting an assistant act across different businesses while preserving the buyer’s understanding and control. Zalando’s experiment could make that future feel normal. It has not yet proved that the difficult part has been solved.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity and the social impact of emerging technology. LiveAIWire is an independent, human-led technology publication using AI-assisted research, editorial production and original AI-assisted editorial illustrations under his direction.