This paper is interested in both the sustainability and scalability of diffusion of pedagogical innovations in educational contexts. We argue that the diffusion of pedagogical innovations is a complex adaptive process. Current research methods, while informative, have limitations insofar as the non-linear, adaptive, and emergent properties of complex diffusion processes are concerned. To this end, this paper proposes the use of Agent-Based Modeling (ABM) as a complementary method. Following a review of existing methodologies, we set up an imperative for ABM, describe what ABM entails, and examine its affordances and limitations in emerging the diffusion patterns by simulating the decision making processes of boundedly-rational individuals in socio-organizational networks. We further argue that ABM is particularly helpful for exploring diffusion patterns in “what if” scenarios, and for making policies that foster sustainable and scalable innovation diffusion.