AI-based process analysis in selection projects and operation of production-related business software

Digitalization challenges small and medium-sized enterprises (SMEs) to support their business processes with suitable software, such as Manufacturing Execution Systems (MES) or Warehouse Management Systems (WMS). However, selecting and implementing such systems is often risky, time-consuming, and expensive. A major bottleneck is process analysis: documentation is frequently outdated or missing, while valuable knowledge about workflows remains implicit in the minds of employees.

The KIProWahl research project aims to significantly reduce this effort. By utilizing an AI chatbot combined with a Large Language Model (LLM), the project develops a solution to conduct automated, context-sensitive interviews with employees in production-related areas (e.g., warehousing, manufacturing). The chatbot captures not only standardized procedures but also the implicit experiential knowledge and individual workflows of the workforce.

Based on these interviews, the system generates real-time process diagrams for both As-Is and To-Be processes. These diagrams are automatically compared and merged to identify deviations and create a consolidated model of company processes. The result provides SMEs with a solid data foundation for software selection or continuous process optimization, reducing the need for extensive methodological expertise or costly external consulting.