The International Association for Automation and Robotics in Construction
Publications / 2025 Proceedings of the 42nd ISARC, Montreal, Canada
Proof-of-Concept Framework of Integrating AI-based LLM Reasoning into BIM Workflows for Design Automation
Building Information Modeling (BIM) has significantly advanced design efficiency and collaboration in the architecture, engineering, and construction (AEC) industry. However, current BIM workflows remain hindered by complex interfaces, repetitive tasks, and limited automation options, which result in a steep learning curve and labor-intensive design processes. To address the challenges, this research presents a proof-of-concept framework that explores the integration of Artificial Intelligence (AI) and Large Language Models (LLMs) with BIM environments to streamline design modifications through natural language interaction. Using a specialized FreeCAD integration layer, the framework demonstrates the feasibility of translating high-level user instructions into basic BIM operations. Initial testing shows this approach can successfully create simple parametric building elements while reducing token consumption and user effort compared to a baseline approach using generic tools. This feasibility study suggests potential for developing more intuitive BIM interfaces, though further research is needed to achieve robust domain knowledge integration and handle complex architectural relationships. The findings indicate promising directions for future development of LLM-driven BIM automation tools that could enhance accessibility for AEC professionals.