Antireflective (AR) coatings are designed to reduce unwanted reflections and enhance light transmission through optical surfaces. They are widely used in lenses, eyeglasses, and display screens. Their performance is quantified by the low average reflectance () over a wavelength range. AR coatings consist of multiple layers deposited on an optical substrate, with input features such as number of layers, sequence of alternating materials, and thickness of each layer. Traditional optimization relies on trial-and-error and extensive experiments, making it costly and time-consuming. In this work, we present an AI-driven optimization approach using high-throughput search to identify AR coatings with  below 0.5% tolerance over a broad range (450–900nm) and narrow range (450–650nm). We trained a CatBoost regression model to map input features to . Shapley value analysis was employed to determine the influence of individual features on . Based on these insights, we generate a dataset of 16 million stacks and identified optimal AR coatings with an =0.04% in the narrow range and =0.28% in the broad range. The recommended coatings were synthesized by our industry partner, Senop Oy, and their superior performance was successfully validated in practice. These results highlight the advantages of AI in industrial manufacturing and product development.