Librarians at San Diego State University (SDSU) leveraged the Primo Search API and generative AI to develop the OneSearch AI Assistant (https://library2.sdsu.edu/onesearch/ai/), an information literacy tool designed to help students translate natural-language research questions into complex Boolean searches compatible with Primo. The assistant takes a question and generates an editable Boolean string that models expert search strategies, making advanced search logic transparent and teachable. This presentation describes the technical architecture of the tool, including prompt engineering, API integration, and design choices made to support instructional goals. While implemented within Primo, the underlying approach is platform agnostic and could be adapted to work with other discovery layers or database interfaces that support Boolean search. The presentation outlines a novel use of the Primo Search API, and shares assessment findings and feedback. Attendees will gain practical insights into developing transferable, instructional AI enhancements for discovery systems.
Prompting Primo Search API and AI
Alma, Primo
Attendees will learn how generative AI and APIs can be combined to scaffold Boolean search skills and how to apply this approach to their own discovery systems.
May 01, 2026 10:00 AM - 10:45 AM(America/Los_Angeles)
Venue : San Gabriel A,B,C
20260501T100020260501T1045America/Los_AngelesPrompting the Primo Search API for Pedagogical Purposes: Building an AI Discovery AssistantLibrarians at San Diego State University (SDSU) leveraged the Primo Search API and generative AI to develop the OneSearch AI Assistant (https://library2.sdsu.edu/onesearch/ai/), an information literacy tool designed to help students translate natural-language research questions into complex Boolean searches compatible with Primo. The assistant takes a question and generates an editable Boolean string that models expert search strategies, making advanced search logic transparent and teachable. This presentation describes the technical architecture of the tool, including prompt engineering, API integration, and design choices made to support instructional goals. While implemented within Primo, the underlying approach is platform agnostic and could be adapted to work with other discovery layers or database interfaces that support Boolean search. The presentation outlines a novel use of the Primo Search API, and shares assessment findings and feedback. Attendees will gain practical insights into developing transferable, instructional AI enhancements for discovery systems.San Gabriel A,B,CEx Libris Knowledge Days and ELUNA Conference 2026eluna-conf-planning@exlibrisusers.org