Tagging and social networks have come into increasing use in concert with the rise of collaborative and interactive on-line media. The focus of tagging is herein twofold: First of all the plain annotation of existing data by a governing instance in order to increase the semantic content of unstructured data, and secondly the application of such meta-information by a community or a group of like minded users. The information contained in such social tagging reflects the point of view and understanding of the community, presenting a valuable source of information for the discovery of community structure,content and intent. This paper proposes an approach aimed at the use of community based tagging to address problems in link prediction and the discovery of complex user groups in a fleeting and unstructured Web-based environment. The ideas presented in this paper are applied to a real world scenario, and the results show a distinct opportunity in community detection and support. This result will be incorporated into emerging knowledge management systems within Siemens AG in the near future.