This Ph.D. project aims to develop a model-driven Knowledge Management treatment for Design Thinking (DT) workshops. DT is both a cognitive framework and a collaboration format structured into phases that produce diverse artifacts: empathy insights, idea catalogs, solution prototypes. DT workshops foster stakeholder commitment through iterative refinement of outcomes. However, current DT practices lack dedicated mechanisms to capture and structure content objects in the semantic context where they emerged. Consequently, Design Thinking knowledge often remains tacit, content remains untraceable, limiting possibilities for retrospective learning, idea evaluation, and strategic reuse. This doctoral project is organized as Design Science research that will engineer a model-driven approach to DT knowledge capture and management. The resulting artifact will streamline a Domain-Specific Modeling Language (DSML) and a Design Thinking Knowledge Graph. The DSML is intended to mimic the visual experience of ideation boards while preserving semantic distinctions in content and adding machine-readability, ultimately turning the visual designs into Knowledge Graphs. This will enable semantic and procedural traceability providing a core structure for potential DT knowledge management systems that can enable organizations to better govern and reuse innovation practices.