Today's IT infrastructures for knowledge management often build on centralized information systems architectures, e.g., Web servers. Such systems have shown their benefits in many situations where knowledge processes are comparatively rigid and where the value of knowledge to be contributed to such a centralized repository can be easily assessed. Unfortunately, knowledge-intensive work frequently exhibits some different characteristics:– It happens sporadically;– It is not tightly embedded in standardized processes or knowledge structures;– It includes previously unseen domains of interest to the knowledge worker;– It requires a lot of ad-hoc interaction with other knowledge workers;– Its value is hard to be assessed and it will end up in a set of best-practices only after a long time.