How can robotics be developed to the point where even smaller companies without their own automation experts can make use of it? Since July 2, 2026, the newly launched research project “AKIMI” has been dedicated to addressing precisely this question — “Agent-Based AI for Intuitively Reconfigurable Robotics in Small and Medium-Sized Enterprises.” The project kicked off with a kickoff meeting at the Chair of Information, Quality, and Sensor Systems in Production at the WZL Machine Tool Laboratory of RWTH Aachen University, where, together with the Institute for Business Cybernetics (IfU) and a broadly representative industry committee, the course was set for practical development.

The situation is likely familiar to many medium-sized manufacturing and logistics companies: volatile markets, shrinking batch sizes, and a noticeable shortage of skilled workers are putting them under increasing pressure. Traditional robot systems often reach their limits in this context because they require significant programming effort and offer little flexibility. This is where the AKIMI consortium aims to step in and develop a hybrid, agent-based artificial intelligence designed to significantly simplify the setup, programming, and reconfiguration of robotic systems.

A bridge between language models and physical implementation

At the heart of this is the idea of linking language models—that is, large language models—with structured domain knowledge and proven basic robotic functions. Unlike purely text-based AI applications, the system is designed to act proactively: it independently breaks down complex human instructions into comprehensible sub-steps and controls the robot’s kinematics accordingly. This would make it possible for the first time to solve assembly and logistics tasks using natural language—even for employees without programming experience. The project team places particular emphasis on the transparency and traceability of AI decisions. Since nothing can be left to chance in an industrial setting, the AKIMI system continuously verifies proposed action sequences against the laws of physics and defined safety limits.

The WZL Machine Tool Laboratory at RWTH Aachen University is contributing its expertise in intelligent assembly automation, led by the Chair of Information, Quality, and Sensor Systems in Production (WZL-IQS) under the direction of Prof. Dr.-Ing. Robert Schmitt. The institute’s own “Sense – Think – Act” framework serves as the methodological foundation: The researchers aim to integrate sensory perception, AI-supported decision-making, and robotic execution as seamlessly as possible. To ensure that the results are not merely theoretical, the developed models and software architectures will be tested on realistic demonstrators as the project progresses and applied to specific use cases of the participating industry partners.

“Small and medium-sized enterprises need flexible automation solutions that can be customized without in-depth programming knowledge. With AKIMI, we’re drastically lowering the barriers to robotics adoption: The agent-based AI acts as an intelligent assistant that translates complex tasks into understandable steps,” explains Prof. Dr.-Ing. Robert Schmitt, holder of the Chair of Information, Quality, and Sensor Systems in Production (IQS) at the WZL of RWTH Aachen University.

The AKIMI research project is funded by the Federal Ministry for Economic Affairs and Climate Action (BMWK) as part of the Joint Industrial Research (IGF) program, pursuant to a resolution of the German Bundestag. The project participants would like to express their gratitude for the support and the provision of research funding.