Expert systems can potentially be used to reduce the cost of building and maintaining training systems. As expert system technology has grown, several tools have become available for building expert systems. The tools are hosted on low-cost work stations or personal computers and provide varying degrees of sophistication for user interface and knowledge representation.

Before, we can rush out and mass-produce useful expert systems, factors that impact the development of expert systems must be investigated. Some of these factors include determining the steps involved in the knowledge engineering process, the qualification criteria for knowledge engineers, performance and limitations of tools available for developing the knowledge-based systems, and development of strategies for integrating the expert systems in the engineering or product environment. These issues are nontrivial and numerous.

This paper begins to attack the problem by presenting experience gained and lessons-learned from a project that involved building an expert system. The objective of the project was to focus on issues relating to the knowledge engineering process, especially that of knowledge retrieval. Therefore, the expert system was developed to solve a relatively simple problem - determine the cause of the malfunction of a modem system.