The Development of AudEx: An Audit Data Assessment System

Abstract

Tools to assist auditors in making sound, complete, and consistent judgements have become increasingly important due to regulation, litigation, and the desire to increase audit effectiveness. Historically, expert systems have been problem specific and created from scratch. Such systems were challenging to update and had limited adaptability when applied to additional problem spaces. If an adaptable system is possible, then costs and time may be reduced related to production, testing, and implementation of quality audit tools. Increased availability of such tools could be a tremendous benefit to auditors. Using a design science research method, this paper describes the process used to create a fraud risk assessment audit tool by embedding audit rules into a generic expert system shell that is widely used in the medical field. The resulting decision and training aid can assist auditors in making fraud risk assessments. While prior research has addressed many components of the issue, the current system was created to address the current risk environment, and a variety of contexts of interest, as the risk environment changes. The system was evaluated by practicing auditors while completing a series of cases. Evaluation results support the benefits of the artifact when used to assist in fraud risk assessments.

Research Ideas for Artificial Intelligence in Auditing: The Formalization of Audit and Workforce Supplementation

“…When applied to auditing, an effective expert system brings numerous benefits such as automatic understanding of audit task processes, as well as increased knowledge and knowledge transferability (Omoteso 2012;Lombardi and Dull 2016). The application of an expert system to accounting, auditing, and tax domain started in the 1980s (Michaelsen 1982;Dungan 1983).…”

Continuous auditing in public sector and central banks: a framework to tackle implementation challenges

“…We rely on the design science information system research and on the design science research process methodologies (Hevner et al , 2004; Peffers et al , 2007). These research methodologies are suitable to tackle identified organisational problems and have been widely implemented not only in the extant literature on auditing (Majdalawieh et al , 2012; Axelsen et al , 2017; Lombardi and Dull, 2016) but also in other related fields (Margherita, 2014; Brooks et al , 2015; Mukherjee et al , 2021). On closer inspection, in the context of our investigation, we define the stages of the research process as follows: Identification of the problems and motivation of the research: Hevner et al (2004) define a problem as the difference between a stated goal and the current state of the situation.…”

In What Way Does Artificial Intelligence Influences Audit Practice? Empirical Evidence from Southwest, Nigeria

“…According to Raphael (2015), effective use of cognitive technologies will enable the audit process "smarter, more intelligent, and more effective" representing the "future of the audit profession" which will benefit financial statement users. Adept expert systems offer advantages in auditing, such as automatic understanding of audit processes and increased knowledge and know-how transferability (Omoteso, 2012;Lombardi and Dull, 2016). Public accounting firms have invested heavily in the development of expert systems with robust knowledge bases to support a wide range of audit tasks, such as planning, compliance testing and substantive testing, risk analysis, and decision making (Brown, 1991).…”