SUSHI approach to automation: short stories of bulk metadata management workflows with automation and AI

Legacy metadata can be challenging to maintain and update as standards and practices evolve over time. Vendor-supplied information typically requires complex processing for fruitful integration into existing discovery systems. Consequently, libraries must figure out how to manage imperfect metadata while being constrained by limited manpower and resources. In this presentation, we will share about the approach taken at the National Library and Archives Board Singapore. Designed by cataloguers, for cataloguers, our workflows efficiently harmonise and resolve legacy and seller-supplied metadata. Through four short stories across different metadata management use cases, we introduce our management approach and philosophy to scripting and AI in cataloguing workflows to increase the efficiency of our metadata services. We explore how these experiences reinforce why cataloguers’ expertise and human-in-the-loop remain paramount, now more than ever, both in designing tech-assisted workflows and working alongside technology to create, maintain and enhance metadata. Keywords: Metadata management, automation, legacy metadata, cataloguing, workflows