An augmented AI-based hybrid fraud detection framework for invoicing platforms

In this era of e-commerce, many companies are moving towards subscription-based invoicing platforms to maintain their electronic invoices. Unfortunately, fraudsters are using these platforms for different types of malicious activities. Identifying fraudsters is often challenging for many companies due to the limitation of time and other resources. A fully automated fraud detection model can be useful, but it creates a risk of false-positive identification. This paper proposed a hybrid fraud detection framework when only a small set of labelled (fraud/non-fraud) data is available, and human input is required in the final decision-making step. This framework used a combination of unsupervised and supervised machine learning, red-flag prioritization, and an augmented AI approach containing a human-in-the-loop process. It also proposed a weighted center based on the feature importance scores for the fraud risk cluster and used it in the red-flag prioritization process. Finally, the approach is illustrated using a case study to identify fraudulent users in an invoicing platform. Our hybrid framework showed promising results in identifying fraudulent users and improving human performance when human input is required to make the final decision.