Dougherty et al. posit that production of complex innovations requires that ecologies be organized, involving three activities: orchestrating knowledge capabilities, ongoing strategizing to frame and direct continuous innovation, and developing public policy to embrace ambiguity. Our study aims to understand how such ecologies emerge. Based on a longitudinal case study, performed in the context of the Quebec health system, our results suggest (a) that the emergence of innovations in highly institutionalized fields requires an additional activity, namely, working on boundaries to make actors perceive their interdependences (b) some levers that can foster the implementation of the model.