Fraud Prediction and the Human Factor: An Approach to Include Human Behavior in an Automated Fraud Audit

Every year, fraud as a subset of insider threats causes billions US dollar of damage worldwide. We suggest a generic architectural model to unify the classic fraud audit approach with human behavior taking into account the fraud triangle in order to achieve better fraud detection and prevention. The human factor is extensively integrated into the audit as a qualitative component, in addition to the classic quantitative analysis of business transactions that are already being applied as part of the fraud audit. This provides added value because transactions examined by the auditor can be better differentiated and prioritized. It is possible to uncover transactions that are part of a pattern that is not yet known and that would have been left undiscovered using normal means by taking suspicious and non-suspicious human behavior into account. The proposed architecture is implemented using a prototype and is applied exemplary to an SAP ERP system.