Analysis of recurring knowledge transfer situations to derive methodological interventions to improve agile product development using AI-based approaches

• Applicants: Professor Dr.-Ing. Albert Albers; Professor Dr.-Ing. Norbert Gronau; Professor Dr.-Ing. Marcus Grum
• Subject Area: Engineering Design, Machine Elements, Product Development
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 559141885

The current state of research shows that agile procedures and methods are already being used in the development of mechatronic systems and are being researched further. Among other things, the focus is on the agile project team, which is in continuous information exchange and regularly reflects on how it can become more efficient and effective. Since the mechanisms of increasing the speed of knowledge transfer and improving the quality of knowledge transfer through methodological interventions have already been empirically tested, these can be used and investigated to what extent they can be used in recurring knowledge transfer situations in agile mechatronic system development (e.g., in daily scrum meetings). e.g. in daily scrum meetings) can be used. At this stage, it is unknown how this can be supported in an automated way by analyzing knowledge transfer situations and selecting appropriate methodological interventions using AI-based approaches. AI is particularly interesting in this regard because it can automatically identify and generalize relationships based on data from previous use cases and interaction histories with product developers and make them available for prediction. Based on these predictions, promising interventions for compensating individual information and knowledge deficits could then be identified by the AI system, so that the speed of realization of development tasks and further the quality of the products could increase. For example, based on a video recording of a virtual meeting, an AI system could determine that a participant has comprehension difficulties that slow down knowledge sharing. The AI system's recommendation to present a technical drawing to clarify comprehension difficulties leads to a predicted reduced meeting time as well as fewer errors in subsequent product generations. The overall goal of this research project is the automated analysis of patterns in recurring knowledge transfer situations and the derivation of methodological interventions to improve knowledge transfer in agile product development using AI-based approaches. In this context, an improvement is characterized by a high result quality (effectiveness) as well as by the relationship between result quality and transfer time (efficiency), where the transfer time results from the transfer speed and the amount of knowledge to be transferred. The factors known from previous projects that influence the transfer speed of knowledge and the quality of knowledge artifacts are used as a basis. In addition, interventions that increase the speed and quality of knowledge transfers in product development are included.

• DFG Programme: Research Grants