Case Study: OCLC achieves higher WorldCat match accuracy and fewer duplicates with GoSecure AI

Gosecure Brings Ai Capabilities to World’s Largest Database of Library Collections

OCLC, a nonprofit global library cooperative that maintains WorldCat™, needed to improve the accuracy and flexibility of its bibliographic matching process to prevent duplicates and ensure high-quality cataloging. To address this, OCLC partnered with GoSecure to explore AI/ML enhancements and develop a production-ready solution for matching new records against the WorldCat database.

GoSecure built and delivered a custom sequence-to-sequence probabilistic NLP model focused initially on audiovisual materials, containerized and exposed via a fast API integrated into OCLC’s existing matching pipeline. The GoSecure solution increased AV match rates by 5–10%, reduced false negatives, ran in under 200 ms per API call, lowered maintenance compared with rule-based approaches, and helped upskill OCLC’s engineering team for future AI work.

OCLC

Ryan Shondell

Executive Director Data Services