Self-verification framework for reliable Text-to-BIM generation accepted at EC3 2026

A paper by the research group of Prof. Christian Bartelt, in cooperation with the research group of Prof. Stefan Lüdtke at the University of Rostock, has been accepted as a short paper at the European Conference on Computing in Construction, EC3 2026, in Corfu, Greece.

The paper, titled “A Self-Verification Framework Toward Reliable Text-to-BIM Generation,” presents an agent-based framework for generating Building Information Modeling, or BIM, models from natural-language prompts. The work addresses a central challenge in applying Large Language Models to engineering design: generated models must not only appear plausible, but also satisfy the technical requirements specified by users.

Recent advances in generative AI have made it possible to create building models from text descriptions. However, many Text-to-BIM systems still rely on users to manually inspect generated models, identify errors, and provide corrections. This limits their practical use in engineering workflows, where reliability, traceability, and verification are essential.

The proposed framework integrates automated self-verification directly into the Text-to-BIM process. For each natural-language prompt, the system derives two complementary forms of validation. First, it generates an Information Delivery Specification, or IDS, for rule-based checking of requirements that can be expressed formally. Second, it uses LLM-driven code-based verification for requirements that are difficult or impossible to represent in IDS. The feedback from both validation steps is then used to iteratively refine the IFC model in a closed loop.

The framework was implemented and evaluated in a case study. The results show that the self-correcting process can improve the quality of LLM-generated BIM models and reduce the need for manual correction. This makes the approach a step toward more reliable AI-assisted design tools for Architecture, Engineering, and Construction workflows.

The work closely reflects the vision behind the new AI Engineering bachelor’s program at TU Clausthal: combining artificial intelligence with engineering expertise to develop reliable, practical, and critically evaluated AI-supported systems for real-world applications.

More information about the AI Engineering bachelor’s program is available here.

The paper website is available here.