Advancing Green Building Certification Through Explainable AI and Automated Data Integration

Green building certification systems—such as Greenstar, GRIHA, LEED, BREEAM, and WELL—play a critical role in advancing sustainable construction and building operations. These systems assess multiple performance dimensions, including resource efficiency, indoor environmental quality, material sustainability, and site impact. However, current certification processes remain complex, time-consuming, and resource-intensive, requiring manual compilation, verification, and interpretation of large volumes of heterogeneous documentation. The increasing adoption of Building Information Modelling (BIM), digital twins, and performance simulation tools has improved data availability and interoperability, yet the certification workflow continues to rely heavily on human expertise to interpret and align technical documentation with prescriptive criteria. This mismatch between digital data generation and manual evaluation constitutes a significant bottleneck in the advancement of widespread green certification.

The research gap lies in the limited integration of artificial intelligence (AI) technologies into the certification process, despite their demonstrated capability in document interpretation, pattern recognition, and predictive modelling. While previous studies have explored AI applications in energy modelling, life-cycle assessment, and construction management, there is a lack of comprehensive frameworks that automate or semi-automate the end-to-end certification pipeline. Moreover, existing AI-enabled tools often focus on isolated tasks—such as extracting building performance parameters or validating single credit requirements—rather than creating a unified system that harmonises diverse data streams and provides transparent, traceable decision support that meets certification authorities’ compliance standards.

This research aims to develop an AI-driven framework that automates key components of the green building certification process by integrating natural language processing, computer vision, rules-based reasoning, and predictive analytics into a coherent, auditable workflow. The objectives are four-fold: (1) to analyse and codify requirements across major green building rating systems into a machine-interpretable knowledge base

(2) to design algorithms capable of extracting, classifying, and validating relevant information from BIM models, performance simulations, and project documentation

(3) to develop an AI-assisted decision-support engine that assesses compliance, identifies gaps, and predicts achievable certification levels

and (4) to prototype an automation tool that generates structured submission materials and provides transparent rationales that align with certification bodies’ review processes.

By addressing these objectives, the research is expected to demonstrate how AI-driven automation can significantly reduce the time, cost, and human effort required for green building certification, while also improving accuracy and consistency. The proposed framework has the potential to enhance sustainability outcomes by making certification more accessible, scalable, and integrated within digital construction ecosystems, ultimately accelerating the transition toward environmentally responsible built environments.