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

Conference

What happens when you combine metadata expertise with AI and use Alma’s bibliographic rank to decide where to begin? At the University of Toronto Libraries, this approach was applied to a long-standing challenge: thousands of government document records too poor in quality and too numerous to address manually. Bibliographic rank, a system-generated indicator of record quality, provided a scalable way to identify records in need of attention, while an AI-assisted Python pipeline enabled the retrieval, validation, and evaluation of candidate records from OCLC to support selection of improved replacements. Once updated, bibliographic rank offered a consistent measure for assessing the impact of these changes across a large dataset. This session presents a practical and replicable approach that combines AI and existing Alma functionality to support metadata remediation at scale, demonstrating how libraries can move from isolated record fixes to programmatic, data-informed improvement.

Development / Initiative