Flexible querying is all about providing users with more intuitive and ecient means of interaction with information or document bases, so that retrieving the items they are interested in becomes easier and more pleasant. In this process three steps may be identi ed as crucial in determining the overall result: understanding the user’s request, understanding the content of what is available and matching between the two. In this paper we present an integrated soft computing framework to tackle all three steps by incorporating fuzzy relational knowledge stored in modi ed ontological structures, a modi ed agglomerative clustering approach to knowledge based information processing, a detailed pro le modeling methodology and a novel approach to the handling of large and sparse transitive relations.