Frankfurt – October 31, 2026 -- Eike Rodenbäck of the German Research Center for Artificial Intelligence (DFKI) won the VDW Research Institute's Research Project of the Year award on September 17 for identifying where AI can pay off for machine tool manufacturers. The prize was presented at the AMB trade fair in Stuttgart for his work in the KI4WZM project.
Three application fields identified for AI in machine tool building
VDW members first flagged their own challenges. Rodenbäck, 28, then identified fields where AI could support the sector economically: knowledge management, fault diagnosis and process optimization. Turning these into concrete pre-competitive projects is planned for a later stage.
On knowledge management, Rodenbäck points to an expected wave of retirements and to manually maintained Excel tables that invite errors and data loss. One project idea is AI-assisted documentation of software code, which would analyze existing codebases and compare them with current documentation to retain know-how when experienced staff leave.
Multimodal diagnostics could cut downtime and service effort
For fault diagnosis, Rodenbäck focused on cases where a problem has already occurred. A multimodal system would link controller alarms with log files, runtime data and machine documentation to find causes faster and provide recommended actions. He says this could shorten downtime and reduce service effort.
Local models can optimize spindle speed and feed without cloud links
AI can also propose parameters such as spindle speed, feed and depth of cut, taking into account material, tool, process data and quality targets. Simpler models that run locally, without external cloud servers, save both money and energy, according to the project findings. Rodenbäck says not every challenge requires a highly complex model.
Workshop draws 60 participants to prioritize follow-up projects
Rodenbäck compiled the results in a detailed report. A workshop for manufacturers organized within the VDW (German Machine Tool Builders' Association) brought the findings to a wider group. The 60 participants also collected, developed and prioritized further project ideas.
Rodenbäck says data quality, reliability, traceability and local data processing remain weak points in most companies, and that these deficits hold back further AI activities.