Identification of tasks for expert systems development in auditing

Abstract

THE ACCOUNTING PROFESSION has shown significant interest in research and development of Decision Support Systems (DSS) and Knowledge-based Expert Systems (KES) in recent years. Reviews of the literature (Abdolmohammadi, 1987; Messier & Hansen, 1987, 1984; O'Leary & Watkins, 1989) indicate that such efforts have centered around complex audit tasks such as analytical review procedures, statistical sampling, auditing electronic data processing centers, going concern judgments, and evaluation of pertinent controls. This is based on some research indicating that expert systems are best suited to tackle unstructured audit tasks (Abdolmohammadi, 1987). Although much time and resources have been spent on efforts to develop expert systems in auditing, these efforts do not represent a systematic approach to identification of tasks for expert system development. Instead, these efforts have been based primarily on consideration of tasks one-at-a-time. This study uses a systematic model to identify viable audit tasks for KES development. The model represents a mechanism to identify audit tasks individually, and in relation with other tasks, within each of the main phases of the audit. The model has the following components: (l) compilation of a comprehensive inventory of audit tasks; (2) organization of audit tasks into audit phases, audit parts within each phase, and audit tasks within each part; and (3) Identification of viable tasks for KES development by very experienced auditors (partners and managers) from international accounting firms that