Digital transformation has revolutionised manufacturing, transportation and finance, yet the construction industry remains one of the least digitised sectors. The lack of automation and information standardisation has led to lengthy approval processes, frequent design revisions, and substantial material waste, limiting innovation and slowing progress. Traditional building design compliance reviews are still conducted manually, resulting in low efficiency, limited transparency, and repetitive effort. This study addresses these challenges by developing a data-driven, intelligent framework for automated compliance checking of building designs. After evaluating various approaches, we determined that combining large language models (LLMs) with computer vision techniques is optimal. For instance, the integrated method significantly outperformed conventional manual and single-technology approaches in both accuracy and processing speed. The proposed framework first translates regulatory texts into machine-readable rules using an LLM, then applies computer vision to extract geometric features and specifications from design drawings. Crucially, a generative, multi-objective optimisation module suggests feasible modifications to non-compliant designs, ensuring the final design meets all regulatory constraints. The optimisation considers multiple objectives simultaneously, producing balanced and actionable design suggestions. This generative revision capability makes the framework not only a compliance checker but also an intelligent design assistant. We validated the system using a dataset of historical building plans from South Australia, representing diverse building types. Preliminary results show the automated system identifies non-conformities with over 90% accuracy and reduces manual review time by approximately 50%. These findings indicate that integrating LLMs and computer vision markedly enhances the efficiency, transparency, and accuracy of building design compliance reviews. By streamlining compliance checks and offering feasible design modification recommendations, our approach ensures building designs satisfy codes and advances the construction industry’s digital transformation. In summary, this proof-of-concept effectively represents a potentially significant step toward digitising compliance enforcement and it supports more innovative, sustainable design processes.