A workflow is a team process since many actors work on various tasks to complete an instance. Resource management in such workflows deals with assignment of tasks to workers or actors. In team formation, it is necessary to ensure that members of a team are compatible with each other. When a workflow instance of, say, an insurance claim (or a surgery) process is performed, the handoffs between successive tasks are often soft as opposed to hard, and actors who perform successive tasks in this process instance must cooperate. If they cooperate well, it can improve quality and increase throughput of the instance. In general, the degree of required cooperation between a pair of tasks varies and this should be captured by a model. This paper develops a model to capture the compatibility between actors while assigning tasks in a workflow to a group of actors. The model is tested through a simulation and the results from a greedy algorithm are compared with optimal results. A technique for computing the compatibility matrix is given and used for an empirical validation from a real execution log. We argue that workflow resource models should recognize soft handoffs and provide support for them.