IGeLU 2026 Conference and Developers Day, 19-22 October 2026

Conference

As Ex Libris integrates generative AI into Primo VE (e.g. Primo Research Assistant), libraries are navigating increasingly opaque, AI-mediated discovery environments. While these tools improve usability, they raise concerns about transparency, oversight, and alignment with established search practices.

This presentation shares a grant-funded project at San Diego State University Library, supported by the Gladys Krieble Delmas Foundation, to develop an explainable AI layer for Primo VE. The prototype translates natural language queries into structured Boolean strings using Microsoft AI Foundry and prompt engineering.

The system sits between user and discovery layer, exposing queries for inspection and refinement. A companion dashboard supports search and surfaces query metadata for transparency. Findings highlight challenges such as known-item searches and indexing mismatches, underscoring the need to align AI outputs with search logic while preserving user control.

IGeLU 2026 theme: "Pearls of Technology: Think deeply, explore widely"
User Experience / Discovery

Primo