Fashion Logistics Robot: Catalogues to ABB AI Picking

A fashion logistics robot does one of the least glamorous jobs in clothing retail. It takes soft, bagged garments out of a pile and puts them one by one onto a sorter, so each order can be built and shipped. This post follows that work from the first mail order catalogues and the huge order plants behind them, through the conveyors and sorters of later warehouses, to online fashion and its returns. It then covers ABB's Fashion Inductor, which gained an AI Robot Trainer in October 2026, the other ABB robots that fit around it, and the parts of apparel handling that are still hard to automate.

ABB's Robotic Fashion Inductor picks bagged items and places them on a sorter. Photo: ABB

How mail order catalogues built the first clothing warehouses

Selling clothes by post began with printed catalogues. Aaron Montgomery Ward mailed his first catalogue from Chicago in 1872, and Sears, Roebuck and Company grew into the largest mail order business in the United States. In Germany, Werner Otto founded his mail order company in Hamburg in 1949. Its first catalogue, in 1950, showed 28 pairs of shoes on 14 hand-bound pages, with photographs of the shoes glued in and a print run of 300 copies. Every order that came back by post had to be read, picked from the shelves, packed and addressed by hand.

Otto's first mail order catalogue from 1950, with photographs of shoes glued onto hand-bound pages. Photo: Otto Group / Wikimedia Commons (CC BY-SA 3.0)

Filling those orders took enormous buildings. Sears began its mail order plant on Chicago's West Side in 1905, and the Merchandise Building was finished in 1906. With about 3 million square feet of floor space, it was called the world's largest commercial building when it was completed, and mail order work went on there for decades. Clerks pulled goods from stock and gathered each customer's items into order bins before packing. Even in a plant that size, nearly every step relied on people reading slips and carrying goods.

Order bins of the Control Delivery Service at the Sears, Roebuck and Company mail order plant in Chicago. Photo: Historic American Buildings Survey via Wikimedia Commons (public domain)

The first big automation in these warehouses moved goods between people. Belt conveyors carried items between floors and departments, and later sortation systems such as tilt tray and cross belt sorters sent each item to the chute for its order or destination. A cross belt sorter carries items on small belts riding a loop and tips each one off at the right exit, which lets thousands of items an hour be split into orders. Someone still had to feed the sorter. Workers stood at induction stations and placed items onto it one at a time, and for clothing that meant picking floppy polybags out of totes and laying each one flat.

A cross belt sorting system in a distribution warehouse, fed by workers at its induction points. Photo: Tecnowey / Wikimedia Commons (CC BY-SA 3.0)

How ABB's Fashion Inductor handles polybags today

Online shopping made that bottleneck bigger. Otto opened an online shop in 1995, and Zalando, founded in Berlin in 2008, built its business on selling fashion online across Europe. Online fashion orders are small, often one or two garments, and many items come back as returns that have to be checked, rebagged and put back into stock. Each of those garments goes through induction at least once. Polybags are slippery and change shape when lifted, so for years this step stayed with people even in warehouses full of conveyors.

Zalando's headquarters in Berlin, home of the online fashion retailer founded in 2008. Photo: Lear 21 / Wikimedia Commons (CC BY-SA 4.0)

ABB launched its Robotic Fashion Inductor in March 2025, together with a Parcel Inductor, as new modules in its Item Picking family. The Fashion Inductor uses AI vision to find a stable grasp point on deformable items such as polybags, picks them from bulk and places them singly onto a sorter. ABB states up to 1,300 picks an hour at peak, with picking accuracy above 99.5 percent. On 15 October 2026 ABB added its AI Robot Trainer and called the result its first end to end vision based physical AI solution. It comes with pre-trained fashion vision models, and ABB says retraining needs no AI specialist.

An ABB robot with a large suction cup, from ABB's 2025 launch of new Item Picking modules for parcels and fashion. Photo: ABB

The trainer runs as a closed loop. It watches robot performance, flags anomalies, generates new training data and sends improved models back to the robot, so picking can keep up when packaging or the product mix changes. ABB says this cuts engineering and commissioning effort for system integrators by up to 75 percent. Cells are designed and simulated in RobotStudio and run on the OmniCore controller, and ABB lists the IRB 1300 and IRB 2600 arms among the related offerings. The module sits beside ABB's Robotic Item Picker for mixed bins, which we covered in our post on parcel sorting and induction.

ABB's Robotic Item Picker, a small ABB arm with a suction gripper above two picking bins. Photo: ABB

Where fashion logistics automation goes next

In warehouses, the next step is to connect induction to the rest of the flow. ABB's Flexley Mover autonomous mobile robots, such as the P604 with 3D Visual SLAM navigation, could bring totes to an induction cell and take finished loads away, so no fixed conveyor has to reach every station. ABB says it will extend the AI Robot Trainer approach to more of its vision based applications through the first quarter of 2027. Linking these pieces is up to each operator and integrator, and nobody has announced a site that combines them.

ABB's Flexley Mover P604 moves loads around a plant without fixed tracks. Photo: ABB

Cost and variety are the open problems. A robotic cell has to pay for itself against shifts of human inducters, which is easier in large, busy sites than in small ones. Fashion packaging also keeps changing, and new polybag films, paper mailers, hangers and boxed shoes all look different to a camera. The ABB throughput and accuracy figures are the company's own stated targets, so buyers will want to test them with their own products. Safety is simpler than in many robot jobs, since cells like this usually work behind guarding, but people still clear jams and refill totes.

The Otto mail order company's headquarters in Hamburg, 2006. Photo: GeorgHH / Wikimedia Commons (public domain)

Home shoppers will feel this mainly as quicker delivery and faster handling of returns. Inside the warehouse, the work itself changes. Staff who once inducted bags by hand will be needed to supervise cells, deal with exceptions and keep the training loop fed with good examples. Integrators will need people who understand both vision models and robot programming, and ABB's decision to make retraining usable without AI specialists is aimed at that gap. Students can practise the robot side in RobotStudio with a no code approach before they work on a real sorter line.

Shoppers in the Zalando outlet store in Berlin, where surplus fashion stock is sold. Photo: Lear 21 / Wikimedia Commons (CC BY-SA 4.0)

Innovation

AI solutions for effortless ABB robot control.

Automation

Robotics

ceojohntran@chatwithrobot.net

+84905311611

© 2025. All rights reserved.

qtran1215@gmail.com