From Tacit to Explicit: Using Live Documentation and Feedback Loops to Facilitate Knowledge Transfer

The preservation and structured use of tacit knowledge is a critical challenge in industrial environments contending with increasing automation and a skilled labor shortage. The loss of undocumented expertise, especially in circular economy applications such as disassembly processes, threatens process efficiency, adaptability and quality. This paper presents a knowledge management approach that combines industrial engineering methods with Industry 4.0 technologies to capture and integrate tacit knowledge into semiautomated disassembly systems digitally. Taking Fraunhofer IFF’s iDeaR project as a case study, a demonstrator is developed to document and convert experts’ actions during PC disassembly into machine-readable formats. The approach integrates live documentation, feedback loops and digital twins to systematically capture contextual problem-solving strategies, enabling their reuse and continuous learning in technical systems. Tacit knowledge is structured using a dedicated Asset Administration Shell (AAS) submodel, comprising situational context, problem description, solution, guidance and benefit. This facilitates contextual reuse across diverse scenarios. The demonstrator architecture links captured knowledge with product, process and resource twins and provides context-sensitive support through modular software applications. Expert-reviewed feedback loops transform raw data into validated disassembly instructions, checklists and training content. A user-friendly interface facilitates intuitive data entry and practical applicability in industrial settings. Results from a workshop-based analysis of disassembly steps confirm that both implicit and explicit knowledge can be meaningfully structured and evaluated for automation capability. The approach preserves expertise, enhances organizational learning and contributes to more adaptive, error-resistant processes. Future developments include AI-assisted storytelling and enhanced sensor integration to further improve feedback quality and reduce editing. This paper thus contributes to the design of intelligent knowledge systems for (semi)automated environments and highlights the value of digital knowledge models in industrial transformation.