This article explores a novel AI‑driven engine that combines large language models with a dynamic knowledge graph to auto‑recommend the most relevant evidence for security questionnaires, boosting accuracy and speed for compliance teams.