SEARCH AI: Enhancing Discovery Through Natural Language

Stony Brook University Libraries is advancing academic library discovery with the development of an AI-powered search tool designed to enhance how users engage with the library catalog. This enhancement, called SEARCH AI, is the first of its kind in a research library environment, allowing users to interact directly with the full library catalog using natural language queries.

Built on an agentic AI framework, SEARCH AI allows users to conduct searches using natural language (phrases and questions more akin to how we speak) rather than traditional keyword-based searches. SEARCH AI parses the user input, identifies its key components, and dynamically outputs a refined Boolean query with relevant filters. Our goals are to provide more relevant search results that help our users find what they need more efficiently and improve usability for all patrons.

We’re currently in the testing and iteration phase, working closely with colleagues to evaluate performance and collect valuable feedback. This grassroots initiative advances artificial intelligence services within a rapidly evolving digital landscape, aligning with the University Libraries’ Strategic Plan, and represents an important milestone in strengthening the Libraries’ AI capacity.

While SEARCH AI is currently being designed for our own library catalog, early results suggest it is adaptable for use at other academic libraries, opening the door to collaborations with institutions that share our commitment to discovery innovation. In the spirit of open scholarship and transparency, the team plans to publish the tool as open-source software on GitHub, accompanied by a full technical write-up.

SEARCH AI reflects our commitment to advancing library innovation, improving discoverability, and contributing to the evolving landscape of academic research.

Stay tuned for updates as development progresses — we’re eager to share more soon.