This paper looks at ways of weaving links between heterogeneous ecosystems, in order to collectively bring about the emergence of new ecosystems at the frontiers of those that initially exist. More specifically, it looks at how to manage the unknowns likely to affect these collaborative ventures (unknowns linked to the objects of joint exploration (in particular the future ecosystems to be explored) and individual unknowns, linked to the individual interest that the joint project will represent for each partner (distribution of the effort to be invested and distribution of the benefits), in situations of socalled “weak orchestration”, in which none of the partners is able to propose what the common object(s) of exploration should be, nor what the terms of collaboration between the partners might be. In this context, we have carried out a research-intervention aiming to observe the logic of action of a particular consortium. This research highlights an original action model for establishing partnerships in weak orchestration situations, which we call Common Lock Quick Learning projects. We characterize the logic that justifies actors getting involved in this type of projects in the unknown. We also highlight the methods required to implement this demanding type of model.