Understanding the possible transition pathways of the energy system requires the integration of human behaviour in energy system models. In order to model the influence of actor behaviour we have developed ACT (Agent-Based Model of Critical Transitions), an agent-based model inspired by an existing conceptualisation of critical transitions. ACT allows us to depart from the current mean-field approach and explicitly explore the effects of heterogeneity, leaders, and networks on the transition. Two key finding are (1) the importance of local communities and (2) leaders can both encourage and discourage the energy transition; a finding that nuances existing literature on critical transitions. We conclude with a reflection on the strengths and weaknesses of our modeling approach.