Siemens and XXL Mechanika Mérnöki Iroda Kft. have partnered to integrate the AI-based Pick AI bin-picking solution into real industrial settings. According to a Siemens case study, the system achieves 97% picking accuracy and 100% bin-emptying efficiency, simplifying a task that previously required complex, part-specific programming and CAD-based recognition.
What changes on the shop floor?
Picking a specific item from a box of randomly mixed shapes and sizes is trivial for a person but has historically been challenging for robots, which needed detailed programming and product-specific recognition. Pick AI uses a pre-trained deep learning model and a 3D camera: the camera analyzes the box contents in real time, identifies objects, and computes optimal, collision-free grasp points in milliseconds.
Boros Gábor, lead electrical engineer and software expert, describes the system as the robot’s ‘‘eyes’’ and decision-support unit: the camera surveys the contents, recognizes items, and determines safe grasp points for the robot.
XXL Mechanika’s role and system integration
XXL Mechanika acts not only as an applicator but also as the Hungarian system integrator for Siemens SIMATIC Robot Pick AI, participating in software testing and development. The company’s engineers assess specific tasks, design tailored solutions, perform programming, and align the robot–software integration with existing production processes.
Kordás Imre, mechanical design and project engineer, emphasized that direct collaboration with Siemens development teams allows them to feed real industrial experience back into the technology and deliver turnkey, complete machine systems. XXL Mechanika intends for Pick AI to move beyond demonstrations and become part of daily production.
Operational flexibility and benefits
A practical advantage of Pick AI is that it can be commissioned without explicit teaching of inventory items, tolerates changes in the parts mix, and can be operated without robot- or PLC-programming expertise when adjustments are needed. The software aims to completely empty a bin; if it cannot pick a piece it will retry or choose another item. It is compatible with various robots, cameras and grippers, allowing delivery of the hardware combination best suited to each task.
Pick AI will be demonstrated at ABB Robotics’ Intralogistics and Automation Solutions Open Days, held in Győr on October 13–15, 202X. XXL Mechanika plans to integrate six to seven ABB robots into its systems next year.
Economic impact and return on investment
The objective of the automation is not to displace workers but to automate monotonous, physically demanding, or hard-to-fill roles so employees can focus on higher-value tasks. Continuous 24/7 robot operation provides consistent quality and does not suffer from human limitations, Kordás noted.
As an example calculation, a typical robotic installation costing around €90,000 with annual labor savings of about €45,000 can pay back in roughly two years. The technology is especially relevant for automotive manufacturing, food processing, e-commerce logistics, pharmaceuticals and warehouse order picking.
Conclusion
The Siemens–XXL Mechanika cooperation aims to make Pick AI a routine industrial tool, offering faster, more flexible bin-picking with reduced need for bespoke programming. With reported figures of 97% picking accuracy and 100% bin-emptying efficiency in a Siemens case study, the system is positioned as a practical response to labour shortages and a way to shift human work toward higher-value activities.
