In this paper, we present the methodology and architecture of the natural language processing integration into all stages of the information systems development. We show that if the IS textual documentation is preprocessed and integrated into the business knowledge base development then the whole information systems modeling process can be speeded and improved. Self-organizing map received from information systems documentation and the formal concept analysis are suggested to test the IS documentation comprehensibility and reusability. IBM’s Information Framework (IFW) Financial Services Data Model (FSDM) has been used for the present research. By using FSDM we demonstrate that the IS model can be partially recreated from IS textual documents by combining techniques based on self-organizing map and formal concept analysis. Finally the numerical experiment is provided to show that IS documents supplemented with the suggested techniques can be reused in natural language interfaces and save the resources and time needed to develop such interfaces.