Existing group awareness system do not scale well with large distributed teams. This article proposes a new way to tackle the information overload problem that arise in such a case. We propose to adapt the aggregation, the filtering and the representation of group awareness information to the user?s work context. The expected result is a group awareness system more pertinent and less intrusive. This adaptation is based on a functional architecture build from a cognitive model of group awareness. This paper explains the foundation of the functional architecture, and how a group awareness tool is made more pertinent and less intrusive by using context to adapt group awareness information.