It is now acknowledged that organizations (companies, institutions, administrations or other) take part into collaborative situations to face globalization, crisis situation, market demand [1] or any other unstable environment. As a consequence, they are strongly dependent on their collaborative processes (and their matching workflows): they have to be efficient and relevant through time. Considering the instability of the collaborative environment (strengths and weaknesses evolve, goals change, partners leave or join the collaboration), agility of collaborative workflows becomes an unavoidable requirement. This agility requirement for collaborative workflows can be expressed as follows: 1) how to detect the instant where the collaborative workflows do not match with the current situation (and what are the causes); 2) how to redefine the best possible collaborative behaviour (at a concrete level, the best workflows and their orchestration and choreography) at time t, in a very short time (close to real time for specific contexts like crisis situations), depending of the current situation, the collaborative workflows progress, the state of the involved resources and collaborative partners. In the remainder of this article, crisis response management domain will be the common theme for the presented examples. Crisis situation response is a specific case of collaboration among organizations in an unstable environment. Their context of execution is moving, unsure and highly sensitive to their evolving environment (a tsunami after an earthquake, pyroclastic flows after an eruption, pandemic into a crowded refugee camp, etc.). In addition, the network of stakeholders involved into the response is also evolving (arrival/departure of partners, denial of service, etc.) and the execution of activities can fail. Plus, the time allowed to react on changes is very short by the nature of the objectives of such collaborations. Moreover, it is necessary to underline that the human factor can also dramatically slow down the process in analysing, combining or even exchanging data during disasters. [2] underline that in the Fukushima disaster management, the main issue, far beyond the technical problems caused by both the earthquake and the consequent tsunami, are the information avoidance behaviours that are related to personal and organizational cultures. This phenomenon was also observed in many major natural or industrial disasters according to [3].