Introduction: The modern construction industry faces the challenge of managing vast volumes of fragmented data, which complicates its integration and analysis across all stages of the construction lifecycle. The use of artificial intelligence, particularly large language models (LLM), offers a promising approach to address these issues. The relevance of the study lies in the need to optimize data processing workflows using LLM to improve the efficiency of construction project management. Materials and Methods: The study examines three methods for applying LLM: Retrieval-Augmented Generation (RAG) for unstructured data, Text-to-SQL for structured data in relational databases, and Knowledge Map for processing sensitive unstructured data. Results: Testing demonstrated that RAG effectively retrieves information from unstructured sources, Text-to-SQL ensures precise access to data in relational databases, and Knowledge Map minimizes errors when handling hierarchical data structures. These methods complement each other, enabling LLM to work with a wide range of data types and improving the accuracy of the responses. Conclusion: The application of LLM in construction data management opens new opportunities for automation and improving the precision of data processing. Further integration of LLM with advanced technologies, such as digital twins and the Internet of Things, has the potential to create more efficient systems for managing construction projects.