AI-Powered Picking Automation at MS Direct: Sereact Enables Robot Without Training

Artificial intelligence is transforming warehouse automation. One of the biggest challenges in robotic picking has long been the ability to handle a large variety of products without extensive programming or training. At the Swiss logistics provider MS Direct, this challenge has now been successfully solved. By implementing Sereact’s AI-powered picking robotics, the company has automated order picking at an AutoStore port for the first time — without the need to train the system for individual items.

Today, the system independently processes around 1,500 orders per day and even operates during night shifts, helping MS Direct increase efficiency, reduce manual workload, and scale operations more flexibly.

MS Direct: A Leading Fulfillment Provider in European E-Commerce

MS Direct AG is a major fulfillment and logistics provider headquartered in St. Gallen, Switzerland. Since its founding in 1978, the company has supported online retailers with scalable solutions in e-commerce fulfillment, cross-border logistics, and returns management.

With more than 600 employees across several locations in Switzerland, Germany, and the United Kingdom, MS Direct processes over 32 million items annually and handles around 17 million product returns each year. The company offers a wide range of services including storage, order picking, shipping, returns processing, and digital logistics solutions tailored to the needs of international e-commerce brands.

Large SKU Variety Made Automation Difficult

As a third-party logistics provider, MS Direct operates several high-throughput fulfillment centers across Switzerland and southern Germany. At its largest facility in Arbon, Switzerland, an AutoStore system from Kardex manages more than 60,000 different SKUs.

This massive product variety posed a major challenge for further automation in the picking process. Many items differed significantly in size, packaging, and surface materials. A previously tested picking robot struggled with these unknown objects, resulting in insufficient flexibility and performance.

As a result, picking operations were still largely performed manually at eight of the nine AutoStore ports. In line with its ambition to become a leading technology player in the European e-commerce sector, MS Direct began searching for a solution capable of handling a wide range of products without complex training procedures. The goal was to automate simple, repetitive tasks, optimize resource allocation, and enable productive night-time operations.

Sereact’s AI-Based Picking Solution

The collaboration with Sereact began through an existing industry contact. Luca Graf, CEO of MS Direct, had previously been in exchange with the Stuttgart-based robotics software developer. During an evaluation phase, MS Direct compared several solutions available on the market before selecting Sereact’s system.

A key factor in this decision was the potential of Sereact’s AI-based control technology to process new and previously unseen products without prior data collection. This technological advantage positioned the solution ahead of many conventional robotic picking systems.

The implemented solution is a pick-and-place system that automates one of the nine existing AutoStore ports. Using advanced visual analysis, the software evaluates each object in real time and identifies key characteristics such as shape, material, and color. Based on this information, the system automatically selects the most suitable gripping strategy.

At the core of the technology is an AI-driven Vision Language Action Model (VLAM) that enables the robot to apply different picking strategies without manual training.

Rapid Implementation and Strong Partnership

The implementation process proved to be remarkably fast. Within just a few weeks, the system was installed, integrated into the warehouse environment, and fully operational — just in time for the start of the peak season.

Marc Messerli, Vice President Fulfillment at MS Direct, emphasizes both the speed of deployment and the collaboration with the Sereact team.

He notes that what impressed the company was not only the technology itself, but also the collaborative and solution-oriented approach of the Sereact team. The partnership was characterized by flexibility, innovation, and a shared enthusiasm for developing new solutions rather than relying on rigid standard processes.

24/7 Productivity: Robot “Pico” Joins the Team

Since the system went live, the picking process at MS Direct has changed significantly. The robot now autonomously processes around 1,500 single-pick orders per day.

The biggest efficiency gains occur during night operations. Previously, the AutoStore system was primarily used during daytime shifts. Now, the robot prepares orders overnight for the next working day, increasing overall throughput while reducing staffing requirements during off-peak hours.

The automation has also achieved another important goal: repetitive and simple tasks have been successfully transferred to robotics. Employees can now focus on more complex picking processes that require human experience and decision-making.

In addition, the system has helped reduce picking errors, minimize failed grasps, and improve overall process reliability within the AutoStore environment. From a financial perspective, the project has also proven highly successful. The return on investment was achieved after approximately nine months, a benchmark supported by the relatively high labor costs in the Swiss market.

To integrate the robot into the company culture, MS Direct even organized an internal naming competition. The robot is now called “Pico” and has quickly become a recognizable part of the team.

Expansion Plans for Automated Picking

The introduction of Sereact’s robotic picking technology represents an important milestone in MS Direct’s automation strategy. The system has demonstrated its reliability in daily operations and proven its ability to process a wide variety of products without prior training.

Encouraged by these results, the company is now planning a gradual expansion of automated picking at additional AutoStore ports at the Arbon site.

Automation is also being considered for other MS Direct logistics locations. These facilities are currently at different stages of technical development, and once the necessary infrastructure is available, the integration of robotic picking solutions will be evaluated.

Beyond picking automation, MS Direct is also exploring additional technologies with Sereact. One option under discussion is the AI-based monitoring platform Lens, which could enable more precise inventory tracking and improved quality control in the future.

With Arbon positioned as the automation pioneer within its network, MS Direct aims to gradually extend these technologies across its logistics operations and further strengthen its position as a technologically leading fulfillment provider in Europe.

For more information, please visit: https://sereact.ai

• wp_statistics_words_count: 976