Detecting Deviations Between External and Internal Regulatory Requirements for Improved Process Compliance Assessment

Abstract

In order to assure process compliance, a wide range of regulatory requirements from various documents must be considered. These external requirements are typically transformed into internal requirements such as policies or handbooks for process compliance in an organization. The transformation is mostly done manually, without the ability of a digitalized quality check. To support users, this work provides a semi-automatic approach based on state-of-the-art NLP algorithms. We first provide a list of Regulatory Compliance Assessment Solution Requirements (RCASR) based on which deviations between external and internal textual requirements can be detected and the root cause of the deviations can be identified. This detailed analysis helps to find mitigation actions in order to improve process compliance. The proposed approach is evaluated based on two Case studies with greatly varying regulatory documents and their realizations by companies. The evaluation demonstrates the feasibility of the approach and provides further insights into the applicability of NLP-based automation techniques in the field of process compliance assurance and management.