Expert Knowledge Elicitations in a Procurement Card Environment

Abstract

Internal Anomaly Detection in the area of employee procurment cards continues to be an issue within many firms, even large multinational corporations. In this case, the authors observed the manual procedures that were being employed by a firm and are creating detailed scripts imitating and enhancing these processes. This firm oversees over 50,000 employee procuremnt card transactions per month. There are over 55 attributes for each transaction. The data set begins in 2011 and is continually updated at the end of each month. Meanwhile, the firm continues to employ its own current manual fraud detection procedures, as a control setting for this study. Based on initial preliminary results, these scripts are expected to improve the success rate of anomaly detection for this firm, plus greatly ease and accelerate its fraud detection process.This research is important as it contributes to the literature of anomaly detection in continuous auditing and monitoring. By replicating the human processes and judgement procedures on many levels, these scripts can improve the audit detection rate at many firms in the area of employee procurement cards.