Purdue University Libraries and School of Information Studies

The Purdue University Libraries and School of Information Studies will design a multi-agent architecture in which coordinated artificial intelligence agents analyze metadata records, identify gaps, retrieve authoritative vocabularies, generate standards aligned enhancements, validate schema compliance, and document provenance and confidence measures. The system will be prototyped across four data ecosystems, including Purdue University Research Repository, Illinois Data Bank, Amazon Web Services Open Data, and National Aeronautics and Space Administration Open Data. The project seeks to transform how libraries convert records into actionable knowledge by developing a multi-agent intelligence infrastructure that prepares metadata and digital collections to be AI-ready and agent-ready for emerging discovery, research, and service ecosystems. This project will focus on establishing a multi-agent architecture composed of coordinated artificial reasoning agents. These agents will analyze existing records, identify metadata gaps, retrieve authoritative vocabularies and ontologies, generate standards aligned enhancements, validate outputs against established schemas, and document provenance, confidence measures, and decision traces.