LAIRA: AI library assistant with multi-API integration and agent reasoning
TL;DR
How academic libraries can incorporate AI-based applications to expand service access while maintaining professional standards is discussed, and the ethical implications are discussed.
Abstract
This study discusses the design and deployment of the Laira (Library AI Research Assistant) at the Library and Learning Resource Centre – RCSI Medical University of Bahrain (RCSI Bahrain Library), outlining the system’s design rationale, implementation process, evaluation findings, and operational challenges. Laira is a reference service that connects end-users to library services and relies on a large language model to support complex information retrieval. The service has been integrated with the library’s discovery platform (Primo), Springshare applications, and a curated institutional knowledge base. Using natural language processing algorithms, Laira identifies relevant sources to respond to user queries. Its main functionalities include dynamic follow-up prompts, multilingual interaction, and real-time service availability. Based on project documentation and empirical implementation experience, the study explains the design development process, the problems encountered during development, and the ethical implications. The study concludes by discussing how academic libraries can incorporate AI-based applications to expand service access while maintaining professional standards.