Abstract Energy efficiency is a critical objective in the Architecture, Engineering, and Construction (AEC) industry, particularly in light of the ongoing energy crisis in Europe. One promising avenue for energy optimization is the building’s façade, with adaptive façades offering significant potential to contribute to the energy grid. Advancements in technological hardware, coupled with the capabilities of Artificial Intelligence (AI), more specific machine learning models, present new opportunities for enhancing façade performance through intelligent control systems. This study proposes a theoretical framework for integrating machine learning models into adaptive façade systems, building upon a patented design. Adopting a design-based approach, the research outlines the formulation of the façade system, including algorithm development and deep reinforcement learning techniques that establish a direct link between the façade’s dynamic adjustments and the needs of a complex environmental context. At the conceptual design stage, recognizing the potential of machine learning in managing complex architectural systems—such as adaptive façades—can lead to innovative solutions that would otherwise be too intricate to operate manually. Furthermore, this study highlights the critical role of data collection in AEC, emphasizing that the limited availability of structured datasets remains a significant barrier to advancing AI-driven architectural applications. Addressing this gap is essential for enabling future research and real-world implementation.
Pantilimonescu et al. (2025) studied this question.