This study examines the effect of strategic prompt engineering on the diagnostic accuracy of generative AI for predictive maintenance in aviation and evaluates the performance of hybrid human AI maintenance teams. Using a mixed methods approach combining quantitative analysis of diagnostic accuracy and mean time to repair (MTTR) metrics with qualitative interviews of industry experts, results show a >11% improvement in fault prediction accuracy and a significant reduction in MTTR. Findings contribute a KPI driven framework and propose governance guidelines for ethical AI deployment in aviation.