Systematic Review of NLP-Driven Approaches for Compliance Gap Analysis in Regulatory Submissions

This paper is a literature review on how gaps in regulatory disclosures can be detected with the help of NLP techniques in 15 studies between 2019 and 2025. It has a discussion regarding transformer-based techniques, domain-based programs on NLP, and check programs on compliance in order to identify gaps in regulation. It has impressive development in automatic checking to achieve a level of 85-96% accuracy and a range of reduction in human review time of 65-85%. Open questions such as interpretability, complexity across jurisdiction, and deployment still remain. Research studies regarding explainable regulation-based AI, transfer learning, and common frameworks still require to be performed. It is a preliminary work in order to create a successful system in NLP to achieve regulatory compliance.