This paper discusses a methodology for multisuategy data analysis based on the application of diverse teaming and discovery programs and toois and how it approaches some of the difficulties posed by the knowledge discovery task. Research in the area of integrated learning systems has led to the develolnent of INLEN, an intelligent assistant for discovering knowledge in large databases. The architecture of INLEN is based on the interaction of a number of knowledge generation operators manifestations of diverse learning tools within a uniform environment. Examples of the system’s application to databases consisting of world economic and demographic facts demonsstrate its operation. During its development, INLEN has encountered problems inherent in the application of symbolic learning progrmns to database analysis that do not appear in the laboratory environment; such problems are described, and the responses to these problems that have been built into INLEN are discussed.