Digital twins have predominantly targeted physical assets, while the ‘human in the loop’—and leadership in particular—remains underexplored 12. We propose an AI-powered Leader Digital Twin (LDT) that models a leader’s memory, routines, and decision heuristics to (i) delegate routine decisions under policy constraints, (ii) preserve and disseminate organizational memory, and (iii) provide what-if analysis for complex scenarios. The LDT integrates five key architectural layers: a data ingestion layer ensuring privacy compliance 3, an organizational knowledge graph encoding people, roles, processes, decisions, and policies 4, a hybrid behaviour model combining LLM-based reasoning with imitation learning 5, a governance and guardrails layer implementing oversight and drift detection 6, and a Decision Theatre UI enabling human-AI collaboration 7. Multi-perspective evaluation protocols, including shadow-mode pilots and NASA-TLX workload metrics, demonstrate promising scalability, robustness, and usability 8. This work addresses critical gaps in leadership succession planning and organizational memory preservation.
Tosi et al. (Wed,) studied this question.