A Comprehensive Review of Predictive Analytics in Audit Risk Assessment for Advancing Data-Driven Decision-Making and Financial Transparency

Abstract

Audit risk assessment remains a foundational element of the financial statement audit, yet traditional approaches that rely primarily on professional judgment and sampling face growing limitations. Rising volumes of transactional data, increasing complexity in financial reporting, and more sophisticated forms of fraud and misstatement have reduced the effectiveness of conventional risk assessment methods. Although predictive analytics, including machine learning, anomaly detection, deep learning, and generative artificial intelligence, has been proposed as a means of addressing these challenges, existing research remains fragmented across techniques, applications, and institutional contexts. As a result, it is difficult to determine how predictive analytics advances audit risk assessment, under what conditions its benefits are realized, and where important constraints persist. This review addresses the problem by systematically synthesizing the relevant literature. It examines the predictive techniques employed, the audit activities in which they are applied, the conditions that shape their effectiveness, and the barriers that limit implementation. Building on this synthesis, the review develops the Integrated Predictive Analytics–Audit Risk Assessment Framework, which organizes technical capabilities, operating mechanisms, moderating conditions, and outcomes related to audit quality and financial transparency. By clarifying the current state of knowledge and providing an integrative structure, the review offers a foundation for more coherent research and practice in data-driven audit risk assessment.