Collaborative virtual environments are becoming an intrinsic part of professional practices. In addition to providing collaboration support, they have the potential to collect vast amounts of multimedia data about the actions and content of such collaborative activities. The aim of this research is to utilize this data effectively, extract meaningful insights out of it and feed discovered knowledge back into the environment. The paper presents a framework for integrating multimedia data mining techniques with collaborative virtual environments, starting from early conceptual development. Discovered patterns are deposited in an organisational memory, which makes these available within the virtual environment. Some of the ideas are illustrated by an example from the application to collaborative spaces developed in LiveNet, a virtual workspace design system.