This chapter discusses the challenges, complexities, and benefits of using agent-based modelling in the study and practice of leadership in teams and organizations. After acknowledging these challenges, the chapter proposes a formal approach to defining leadership as an influence process for computational models and simulation purposes as well as in real-world organizations. This is intended to clarify how leadership can be specified, modelled, and operationalized more rigorously in the context of both research and practice. State-of-the-art studies in computational leadership research, including those with adaptive agents, adaptive environments, and emergent intelligence, are reviewed. These studies are organized and described along three distinct dimensions of leadership influence: aligning expectations in a collective with respect to opportunities and threats in its environment, initiating or sustaining repeatable instrumental momentum within organizations, and engendering an observable cooperating valence within a collective that reflects the presence of trust among agents at various levels of scale. The chapter closes with a proposed framework to guide future research in areas such as emergent leadership in hybrid human–machine environments and decision-making in the presence of localized deception during interactions as well as the potential to respond effectively to misinformation attacks from malicious actors on a much broader scale.