Data-Driven Audit Intelligence: Enhancing HSE Assurance Through Automated Audit Planning and Execution

Abstract This paper introduces a data-driven audit intelligence framework designed to modernize and strengthen the organization's Safety, Security, Health, and Environment (SSHE) assurance process through the integration of analytics, automation, and artificial intelligence. The framework aims to overcome the limitations of traditional manual audit planning, which often relies on subjective judgment and post-incident analysis. In organizations that manage multiple audit programs across numerous operational sites, manual scheduling and checklist preparation tend to create redundancies, inconsistent risk coverage, and inefficiencies in resource allocation. The proposed framework establishes a data-informed, risk-based approach to audit planning transforming auditing into a proactive tool for operational risk prevention rather than a retrospective compliance activity. At the foundation of this framework is a centralized Data Mart that consolidates both structured and unstructured information, including audit findings, incident data, and regulatory requirements. Large Language Model (LLM) techniques are employed to extract entities, insights, and semantic relationships from textual audit reports transforming unstructured narratives into structured knowledge. These extracted insights are integrated with quantitative indicators such as incident frequency, trend direction, and severity levels to generate dynamic risk scores for each operational site or process. The risk scores serve as the input for algorithmic optimization models, where Python-based linear constraints are applied to reflect audit frequency rules, program types, and regulatory mandates. This optimization process automatically generates multi-year and annual audit schedules that ensure adequate coverage of high-risk areas while minimizing redundancy. Additionally, the framework includes the capability to auto-generate audit checklists based on historical findings and recurring nonconformities, allowing auditors to focus on the most relevant verification points. Implementation of the framework has demonstrated tangible improvements in efficiency, coverage, and transparency within the audit process. Heatmap visualizations highlight critical risk zones and trend patterns, providing management with a clear, data-driven perspective of SSHE performance across operational assets. Meanwhile, interactive dashboards deliver real-time situational awareness, enabling audit leads to monitor progress, review compliance trends, and re-prioritize areas as new risks emerge. Auditors report increased confidence in aligning audit scopes with operational risk priorities, while management benefits from a more systematic and traceable decision-making process. Ultimately, this framework establishes auditing as a core component of predictive governance, empowering the organization to anticipate risks, optimize resource allocation, and continuously strengthen its SSHE management system through data-informed intelligence