Safety and health management systems (SHMS) translate legal duties, risk assessments and operational learning into practical work instructions. In mining, they can include hundreds of procedures, standards, job safety analyses, risk assessments, permits, competency matrices and verification tools. These documents are revised at different times by different owners. Over time, document drift can create conflicting instructions, obsolete legislative references, duplicated controls, mismatched inspection frequencies and unclear escalation triggers. These weaknesses increase procedural burden, reduce field usability and make control assurance less reliable. The issue is not only document control; it is a human-centred design problem because workers and supervisors must interpret the system under time pressure and variable conditions. ISO 45001:2018 frames occupational health and safety management systems around legal compliance, worker participation, operational control, evaluation and continual improvement (ISO, 2018). The International Labour Organization also emphasises continual improvement in occupational safety and health management systems (ILO, 2001). This extended abstract presents an artificial intelligence (AI, OpenAI, 2026) supported method for continuous SHMS quality assurance, focused on whether documented controls have coherent logic against unwanted events.