This thesis explores how conversational Artificial Intelligence, particularly Large Language Models (LLMs), can be used to make Building Information Models (BIM) more accessible and understandable. The goal is to enable users to interact with BIM data through natural language questions, without needing specialized technical knowledge. To support this, BIM data in the form of Industry Foundation Classes (IFC) is converted into a structured knowledge graph that represents spatial and semantic relationships within the building. A multi-agent system guides the process of generating and improving database queries based on user input, making the system more reliable and responsive. The result is a web-based platform that allows users to ask questions about a building and receive accurate answers, supported by both text and 3D visual feedback. The system was tested using a dataset of questions collected from professionals in the Architecture, Engineering, and Construction (AEC) industry. This work demonstrates the potential of combining AI and BIM for more intuitive information access, laying the groundwork for future tools that help professionals better explore, understand, and make decisions using complex building models. Key Terms: Knowledge Graphs, Large Language Models, BIM, IFC, Multi-Agent AI, Conversational AI     «

This thesis explores how conversational Artificial Intelligence, particularly Large Language Models (LLMs), can be used to make Building Information Models (BIM) more accessible and understandable. The goal is to enable users to interact with BIM data through natural language questions, without needing specialized technical knowledge. To support this, BIM data in the form of Industry Foundation Classes (IFC) is converted into a structured knowledge graph that represents spatial and semantic rela...     »