This paper reports on an ongoing development of an application that aims at identifying communities of use in the context of an organized group of people. The described application, called CoWing consists on mining users bookmark files in order to identify communities that share the same information interests. The system is composed of a set of assistant agents, called WINGS, and a central agent that manages the users organization. A Wing agent observes the user behavior in order to learn the user bookmark classification strategy. A hybrid neural case-based reasoning incremental classifier is used for this purpose. Classification knowledge learned by an agent is then used to identify, for each folder in the local bookmark hierarchy, a community of users that share the same interests in the theme of that folder.