A conceptual framework for the network approach to expert systems development in auditing

Abstract

The process of audit opinion formulation is complex. Before the final opinion decision is reached, the auditor must make a large number of related decisions which contribute to the final opinion decision. We analyze this complex audit decision process as a special case of a more general decision process model. The results of the analysis are used to classify and analyze existing expert systems in auditing. It is clear that the systems proposed to date do not deal with the audit process as a whole, but rather with more limited decisions within an audit. In order to adequately address the complete audit, we propose a network of expert processors. Each processor will deal with a specific decision problem within the context of an audit. However, since the processors can communicate with each other, the separation of the tasks does not adversely affect the aggregation of the information during the audit. The network concept will allow expert systems researchers in auditing to pursue relatively narrow domains of audit knowledge with the assurance that they are contributing to the development of a complete audit expert system.