The energy efficiency of educational buildings is critically influenced by the building envelope, particularly the Window-Wall System (WWS). While the window-wall ratio is a well-studied parameter, a holistic understanding of the integrated WWS remains limited. This study bridges this gap by developing artificial intelligence models to predict the collective impact of nine critical WWS factors-including glazing type, solar heat gain coefficient, shading devices, thermal insulation, and air leakage rate-on the energy consumption of school buildings across extreme climatic gradients (-10 degrees C-40 degrees C) in Pakistan. A substantial dataset was collected from 6301 schools across six cities representing hot desert, humid subtropical, and cold semi-arid climates. Four prominent AI models-Artificial Neural Network (ANN), Extreme Gradient Boosting (XGBoost), Random Forest, and K-Nearest Neighbors-were trained and rigorously validated. The ANN model demonstrated superior predictive capability, outperforming other models and explaining 88 % of the variance in energy consumption. A subsequent Pearson correlation analysis quantified the influence of individual factors, identifying the air leakage rate (r = 0.7005) as the most significant driver of increased energy use. In contrast, glazing type (r =-0.5952), thermal insulation (r =-0.5943), and shading devices (r =-0.5896) were the strongest contributors to energy savings. This research provides a robust, data-driven framework that empowers architects, engineers, and policymakers to make informed decisions for retrofitting existing schools and designing new, energy-efficient educational infrastructures tailored to diverse climatic conditions.