The Journal of Academic Librarianship Volume 52, Issue 4, July 2026 DOI: 10.1016/j.acalib.2026.103255
Abstract
This article presents a case study on the development, phased testing, and user evaluation of Library Information & Knowledge (LINK), a generative AI–enabled library chatbot created by the University of South Florida Libraries. Built using Microsoft Copilot Studio, LINK integrates conversational AI with structured knowledge documents to extend library services, improve efficiencies, and support user navigation of library resources. Development required extensive groundwork, including web content mapping, knowledge base restructuring, and iterative refinement informed by internal staff, faculty, and student testing. An IRB-approved survey captured user perceptions of ease of use, accuracy, usefulness, and trust. Respondents valued LINK’s 24/7 availability and conversational flexibility, but noted limitations tied to its ethical design constraints, particularly its inability to access licensed databases. These findings reflect broader trends in the literature, which highlight the tension between user expectations and the actual capabilities of AI chatbots in academic settings. The study identifies priorities for future development, including reducing the database access gap, expanding knowledge coverage, improving accuracy with structured and source-linked responses, and strengthening pathways to human assistance. Despite technical and structural challenges, the project demonstrates how low-code generative AI tools can enhance academic library service ecosystems while reinforcing transparency and user trust.
Leave Your Comments