Zalando is taking AI out of the chatbot and into the warehouse
24 July 2026
By Keegan Moore – Director of Business Development
Its investment in Sereact points to where AI could make a real difference in fashion eCommerce: fulfilment and returns.
Most retail AI announcements are easy to spot: shopping assistants, recommendations and generated product pages. Zalando’s latest move is different. It is putting money into the part customers rarely see, but retailers feel directly in the P&L: the warehouse.
Zalando has joined the $116 million Series B round of Stuttgart robotics company Sereact as a strategic investor. It has not said how much it invested or when the technology will be deployed. Even so, the direction is clear. This is less about buying robotic arms and more about backing software that can handle less predictable work.
Returns are the hard bit
Fashion returns are messy. A parcel might contain an unopened pair of shoes, a creased dress or an item with damaged packaging. Each product has to be opened, identified, checked, repacked and returned to stock. The longer that takes, the greater the chance that a full-price item ends up discounted.
Traditional automation is good at repeating the same movement. Returns rarely offer that consistency. Sereact uses computer vision and AI-based instructions to help robots respond to different products, materials and packaging rather than follow one fixed routine.
A key portfolio decision
This follows Zalando’s earlier investment in Nomagic, whose robots handle item-level picking and sorting. Together, the investments point to a practical approach: work with specialists solving specific problems rather than trying to build everything in-house.
There is a wider angle too. Zalando increasingly sells fulfilment services to brands and retailers through ZEOS and its B2B business. Better automation could improve its own cost base while making that service more attractive to partners.
What retailers should take from it
The lesson is not that every retailer needs robots. Start with the bottleneck. Where is manual work costly, difficult to staff, inconsistent or slowing inventory down? Automation makes sense when the volume is there and the economics hold up beyond a controlled pilot.
It still depends on the basics: accurate product data, system integration and clear processes for the cases a robot cannot handle.
The broader signal
The first wave of retail AI was largely about content. The next will be judged on operations: whether it gets products back into stock faster, cuts handling costs and makes fulfilment easier to scale. Less glamorous, perhaps, but probably more important.