Introduction: This presentation details an assessment of SEARCH AI, a home grown semantic search tool built using generative AI that empowers patrons to query the library catalog using plain, natural language.

Methods: Three health sciences librarians conducted a comparative study to assess the efficacy of SEARCH AI against traditional catalog searching. The study utilized actual Fall 2025 undergraduate health sciences student research topics to query the catalog. For each topic, a traditional search was compared to an AI-generated search. Data collected included the number of results, and the relevancy of the first ten results to the student's research need. Furthermore, the AI-generated Boolean search strings were assessed for accuracy using a subset of criteria from the PRESS checklist.

Results: Initial findings suggest SEARCH AI has the potential to enhance the discoverability of resources for novice users, transforming conversational searches into more complex Boolean queries.

Discussion: The results will be used to directly inform the development of an integrated library instruction module focused on educating health sciences students on the strategic, critical use of this AI search feature, teaching them not only how to use the tool but also how to interpret its output (specifically the generated Boolean strings) to foster advanced search proficiency—a cornerstone of evidence-based practice.