Evaluating the Impact of Data Analytics on Audit Evidence Collection and Risk Assessment Accuracy

Posted: Apr 14, 2018

• Liam Rivera
• Annalise Barrett
• Theo Watson

Abstract

This research investigates the transformative effects of advanced data analytics on the fundamental audit processes of evidence collection and risk assessment accuracy. Traditional audit methodologies have long relied on sampling-based approaches and manual testing procedures, which inherently carry limitations in coverage and precision. Our study introduces a novel framework that integrates machine learning algorithms with natural language processing techniques to analyze complete populations of financial transactions and supporting documentation. We developed and tested this approach across three distinct industry sectors—financial services, manufacturing, and healthcare—involving over 15 million transactions and 50,000 documents. The methodology employs unsupervised learning for anomaly detection, supervised classification for risk pattern identification, and semantic analysis for evidence validation. Our findings demonstrate a 47