Abstract
The paper presents an agent-directed digital twin framework for simulation-based training realising a synergy between digital twin concept, intelligent agent technologies and simulation gaming. The proposed framework provides a unifying formal basis for robust digital twin implementations with various application possibilities. In this work, the research focus is made on simulation-based training to improve business processes in an organisation. In contrast to typical applications of digital twins in manufacturing and other cyber-physical processes related to product life cycle management, the pre-sented framework highlights relatively new application area of organisational digital twins. The experimental capabilities of the agent-directed digital twin framework are illustrated by a remote work organisation training model.