Smart multimodel containers: AI supported automatic linking

This paper presents a novel framework for AI-supported semantic querying and automatic link generation within ISO 21597-compliant multimodel containers (MMCs). The proposed system enables intuitive interaction with Building Information Modeling (BIM) data through a natural language interface (NLI) that supports non-technical users in exploring, querying, and linking complex construction-related datasets. The framework combines ontology-guided natural language processing, intent classification, and unit-aware querying to dynamically interpret user queries and extract relevant data from Industry Foundation Classes (IFC) models. Furthermore, the system generates semantic links between BIM elements and heterogeneous data sources without modifying the original models. A layered architecture is adopted, integrating modules for ontology mapping, IFC querying, and ICDD-compliant link generation. The framework is validated through an exemplar use case in which users interact with a BIM model and IoT sensor data, demonstrating the system’s ability to retrieve relevant building elements and generate dynamic links to their associated time-series data. Results show that the proposed approach enhances data accessibility, traceability, and interoperability in digital construction environments, laying the foundation for more dynamic and intelligent multimodel container systems.