AI-enhanced reference services in special libraries: a case study of the Hakka Literary Museum

Purpose This study explores integrating artificial intelligence (AI) technologies into reference services within special libraries, focusing on the Hakka Literary Museum as a case example. It aims to address the limitations of traditional reference services in meeting complex and domain-specific information needs, especially under constraints of limited human resources. Design/methodology/approach A case study methodology examines the layered implementation of AI technologies, including retrieval-augmented generation (RAG) and GraphRAG. These tools support four tiers of reference services: directive reference via real-time Web scraping, conceptual reference using RAG for topic-specific responses, deep-indexing through iterative document expansion and thematic reference via GraphRAG-powered hierarchical knowledge graphs. Each layer is tailored to support varying depths of user inquiry. Findings This study finds that AI-enhanced reference services significantly improve service responsiveness, content depth and user satisfaction in cultural and special library settings. The system enables accurate and efficient responses to both basic and advanced inquiries, bridging gaps in human resource availability and enabling continuous access to cultural knowledge resources. Originality/value This research contributes to the growing literature on AI in information services by offering a practical model for implementing multilayered reference assistance in special libraries. It highlights the novel use of GraphRAG for cultural knowledge mapping and demonstrates how AI can support the preservation and dissemination of intangible heritage. The study also provides insights into the transformative role of AI in redefining user experience and operational efficiency in information services.