FRAUD-SCAN AI: An Explainable Early-Warning System for Corporate Fraud Risk Screening

About the Project

Financial fraud remains a significant concern for corporate transparency, investor confidence, and sustainable economic development. Corporate annual reports contain extensive financial information, management narratives, and governance disclosure. However, manually reviewing this complex disclosure is time-consuming, and potential fraud risk signals may remain difficult to identify and priorities. This innovation introduces FRAUD-SCAN AI, an explainable early warning and decision support system designed to transform annual report information into actionable fraud risk intelligence. Building upon research involving Natural Language Processing (NLP), sentiment analysis, and corporate governance analysis, a functional research prototype has been developed to extract and screen annual report information across narrative, governance and financial distress dimensions. Identified signals are consolidated into a prototype Fraud Signal Risk Index (FRSI) and presented through an interactive risk monitoring dashboard. A key feature is its explainability mechanism, which traces identified signals to relevant evidence within the annual report and explains why particular information may warrant further attention. FRAUD-SCAN AI is designed to support auditors, regulators, audit committees, and governance professionals in screening companies, prioritizing higher risk cases, and directing professional attention towards areas requiring further review. Rather than determining whether fraud has occurred, the system functions as an early warning screening tools while preserving professional judgement. Following empirical validation and further system development, FRAUD-SCAN AI has potential for commercialization through institutional licensing, subscription-based platform access, and specialized professional screening services, with opportunities for future adaptation beyond the Malaysia corporate reporting environment. Keywords: Fraud Risk, Annual Report, Natural Language Processing, Sentiment Analysis, Corporate Governance, Explainable AI, Fraud Risk Signal Index