Artificial intelligence (AI) is increasingly combined with Operations Research (OR) to solve optimisation problems across sectors, yet existing reviews study each sector in isolation. This study presents a cross-sector bibliometric review of 1,550 Scopus-indexed articles and reviews published between 2016 and 2026, analysed using Bibliometrix, Biblioshiny and VOSviewer. Publications have grown rapidly, with 88.3% appearing since 2024, and the field is highly collaborative and spread across many journals. Thematic mapping identifies predictive analytics and decision-making as the motor theme, machine learning and optimisation as basic themes, and large language models as an emerging theme. The review finds five gaps: prediction dominates prescription, real-world validation is limited, sectors are weakly integrated, adoption barriers are rarely examined, and fairness and AI costs get little attention. A six-priority research agenda is proposed to address them.