We review systems that support the management of collaborative interaction, and propose a classification framework built on a simple model of coaching. Our framework distinguishes between mirroring systems, which display basic actions to collaborators, metacognitive tools, which represent the state of interaction via a set of key indicators, and coaching systems, which offer advice based on an interpretation of those indicators. The reviewed systems are further characterized by the type of interaction data they assimilate, the processes they use for deriving higher-level data representations, and the type of feedback they provide to users.