This lecture explores AI-driven HVAC optimisation, moving from rule-based automation to adaptive, data-driven control. It covers machine learning for load prediction, occupancy modelling, anomaly detection, demand response, weather-adaptive control, and reinforcement learning, with examples from real-world deployments. The lecture also examines challenges such as data quality, model drift, explainability, and operator trust, highlighting the potential of AI to improve both energy efficiency and occupant comfort.