Energy-intensive sectors such as industry and transportation are vital to economic development, yet remain major contributors to global energy consumption and greenhouse gas emissions. Enhancing their energy efficiency is essential to achieving climate resilience and meeting decarbonization goals. Despite abundant operational and sensor data, decision-making in these sectors often relies on heuristics and lacks systematic, transparent, and explainable analytical support. Existing Artificial Intelligence (AI)-based decision support methods frequently fall short in integrating robust data preprocessing, causal interpretation, and actionable recommendation generation, limiting their practical impact. This book develops and validates an AI-driven decision support framework to improve energy efficiency in energy-intensive industrial and transportation systems. Guided by a three-cycle Design Science Research (DSR) methodology, the framework integrates complementary AI techniques into a modular, end-to-end architecture that transforms heterogeneous raw data into interpretable, actionable recommendations. It includes systematic data quality assessment and preprocessing pipelines, time-series segmentation, clustering key performance indicators for pattern recognition, causal inference methods to identify drivers of inefficiency, and a multimodal large language model (LLM)-based decision support module that translates analytical outcomes into domain-relevant strategies.