How is AI used in fashion logistics and warehouses?
From picking robots to slotting and delivery date promises, AI is moving into fashion fulfilment. What is deployed, what it depends on, and why soft, varied products remain hard to automate.
KEY TAKEAWAYS Summary by the editors
- AI in fashion logistics covers three main areas: robotic picking and handling, warehouse optimisation such as slotting and labour planning, and more accurate delivery promises to customers.
- Zalando uses AI-powered picking robots from Nomagic that pick, scan and induct items into pocket sorters; a pilot averaged 10,000 picks per day, with nine robots live in 2025 and a double-digit number planned for 2026.
- Fashion is difficult for robots because items are soft, packed in polybags and highly varied, which is why computer vision and machine learning matter more than in rigid-goods warehouses.
- Warehouse AI depends on accurate item master data, location data and order history; poor data limits slotting and forecasting gains.
- Automation investment is a long-term bet that must be weighed against volume volatility, returns handling and the flexibility of human labour.
AI in fashion logistics is used mainly for robotic picking and item handling, for warehouse optimisation such as slotting and labour planning, and for more reliable delivery date promises. Real deployments exist, such as Zalando's use of AI-driven picking robots, but they cover specific steps rather than whole warehouses. The value depends on volume, product mix and the quality of item and location data.
Why is fashion logistics hard to automate?
Fashion warehouses handle very large numbers of SKUs in sizes and colours, short product lifecycles, seasonal peaks and high return rates in online channels. The items themselves are a challenge: soft garments in transparent polybags, shoes in boxes of varying size, accessories of every shape. Classic automation works best with uniform, rigid cartons. Handling individual, deformable items needs machines that can see and adapt, which is where AI comes in.
What does AI-powered picking look like in practice?
In October 2025, Nomagic announced that Zalando had selected its robots to expand robotic capabilities in its fulfilment centres. According to Nomagic and Retail Systems, the robots pick individual items, scan them and feed them into automated pocket sorters. A pilot reached an average of 10,000 picks per day; nine robots were to be live in 2025, with a double-digit number planned for 2026. Computer vision and machine learning allow the robots to handle a wider range of products and adapt to changing inventory. Zalando also made a minority investment in Nomagic's 44 million dollar Series B round.
In July 2026, FashionNetwork reported that Zalando had invested in Sereact, a Stuttgart-based developer of AI for robotic picking and returns processing. The report describes the investment as supporting the international rollout of Sereact's technology and notes that Zalando's logistics unit relies heavily on automation; it does not state that Sereact systems are running in Zalando warehouses.
These examples show the current pattern: robots take over a defined, repetitive step, such as picking into a sorter, while people continue to handle exceptions, replenishment and many returns tasks.
Which AI use cases matter most in a fashion warehouse?
| Use case | What it does | Data needed | Maturity |
|---|---|---|---|
| Robotic piece picking | Picks single items from totes into sorters or orders | Item images, dimensions, packaging type | Deployed at scale by some large online retailers |
| Slotting optimisation | Places fast movers and items often ordered together close to pick stations | Order history, item dimensions, location map | Established, increasingly ML-based |
| Labour and wave planning | Forecasts workload by hour and schedules staff and waves | Order forecasts, productivity data | Established |
| Returns grading | Assesses returned items and routes them to restock, refurbish or resale | Images, product data, return reasons | Emerging |
| Delivery promise | Estimates delivery dates per order and carrier | Stock location, carrier performance, cut-off times | Established, accuracy varies |
How does AI improve slotting and delivery promises?
Slotting decides where each item lives in the warehouse. Machine learning models can predict which items will be ordered together and how demand shifts across a season, then suggest moves that shorten picking paths. In fashion, where assortments change constantly, slotting that adapts weekly rather than once a season can make a measurable difference, provided the location and item data are reliable.
Delivery promises are the customer-facing side. A model that combines stock location, warehouse workload and carrier performance can show a more accurate delivery date at checkout. An honest promise matters more than a fast one: missed dates generate customer contacts and, in fashion, can lead to returns when an item arrives after the occasion it was bought for.
Inventory placement links both topics. Deciding which warehouse or store holds which stock, and how much, determines how quickly orders can be fulfilled and how often items must be split across parcels. Forecasting models that work at size and location level can support these decisions, but in fashion they face short histories for new styles and strong effects from weather, promotions and trends. Wholesale and marketplace channels add further complexity, because the same stock pool may serve retail partners, own stores and online orders with different service levels.
What are the limits and risks?
- Capital intensity: robots and sorters are long-term investments that need stable volumes to pay back.
- Volatility: peaks such as sales events and holiday seasons stress automated systems designed for average loads.
- Data quality: wrong item dimensions or missing images reduce picking success and slotting accuracy.
- Returns complexity: grading returned garments for resale still needs human judgement on condition.
- Workforce impact: automation changes roles and skills in warehouses, which needs planning and dialogue.
McKinsey's State of Fashion 2026 lists 'Efficiency unlocked' and 'Workforce rewired' among its ten themes and finds executives expecting difficult conditions, with 46 percent expecting conditions to worsen in 2026. In that environment, logistics automation competes for capital with many other priorities and is likely to be judged on clear operational returns.
What should fashion companies consider before investing?
- Map where time and errors concentrate in your warehouse: picking, packing, returns or replenishment.
- Check data completeness for item dimensions, packaging and images.
- Model volumes and peaks over several years, not one season.
- Pilot on one process with clear metrics such as picks per hour, error rate and uptime.
- Plan staff roles and training alongside the technology.
- Consider whether third-party logistics providers already offer the capability as a service.
What is the outlook for AI in fashion fulfilment?
The direction is towards more flexible, AI-driven robots that can handle a wider range of items and tasks, including returns. Large online retailers are investing both as users and as investors in robotics companies. For mid-sized brands and wholesalers, the nearer-term gains are more likely to come from software, slotting, labour planning and delivery promises, built on clean product and inventory data.
Frequently asked questions
Does Zalando use robots in its warehouses?
Yes. According to Nomagic and Retail Systems, Zalando uses Nomagic's AI-powered robots to pick, scan and induct items into pocket sorters. A pilot averaged 10,000 picks per day, with nine robots live in 2025 and more planned for 2026.
Why is it hard for robots to pick clothes?
Garments are soft, often packed in shiny polybags and vary greatly in shape and size. Robots need computer vision and machine learning to recognise and grip such items reliably.
What is warehouse slotting?
Slotting is deciding where each item is stored so that picking is efficient. AI-based slotting uses order history and forecasts to place fast-moving items and items often ordered together in the best locations.
Is warehouse automation worth it for mid-sized fashion brands?
It depends on volume, stability of demand and data quality. Many mid-sized companies gain more from software improvements or from logistics providers that already operate automation than from owning robots.
One edition every weekday morning. Read in five minutes. Free for industry professionals.
