Discrete-event digital simulation modelling is an important and increasingly used technique for analysing and solving complex problems in business, industry and the public service. Computer-aided systems are under development to improve the efficiency of what is currently an expensive time-consuming process. In particular, problem formulation is slow and inexact, an art which improves with experience. Artificial-intelligence aids have been developed to help the analyst in collaboration with the decision maker to solve the problem. The first attempts were based on expert systems. Limitations with these systems led to the development of a natural language understanding system. The way in which the natural language understanding system undertakes the task of aiding problem formulation is described, as well as the structure of the system itself. The role that such a system plays in the overall modelling process is described, with emphasis on limitations of the current system and how these might be overcome in future developments.