Implementation of machine learning in internal audit : a case study of a multinational Belgian company

Abstract

The objective of this study was to understand the utility of Machine Learning methods in internal audit. It provides a common framework for implementing Machine Learning in internal audit in order to attain Continuous auditing. The research begins with the review of the existing literature, followed by interviews with AI experts and internal auditors, since a team working on such a project would need to have both these profiles. It then takes the case study of implementing AI within a multinational company's internal audit department, focused on detecting outliers in the Expenses Report. The outcome was the creation a Proof of Concept AI model reaching 88% prediction accuracy, an API to access the model and a user interface for the auditors. This study highlights the different iterations of training the model but also the importance of ML techniques in internal audit for the future; AI will radically change the way auditors go about doing their job.