The Technical Services Workflow Efficiency Interest Group (TSWEIG) is pleased to host three presentations during Core Interest Group week.
The program will take place on Monday, March 2, 2026, from 2:00-3:00 pm Central Time.
The session is online, free to attend, and open to the public. To register, please click on the following link: https://www.ala.org/core/interest-group-week
Human-in-the-Loop Automation: Applying APIs, MARCXML, and AI Concepts to Technical Services Workflows / Alvin Stockdale
Technical services workflows often rely on repetitive, manual processes that are difficult to scale across your entire catalog. While many libraries are experimenting with AI tools, meaningful efficiency gains often come from well-designed automation that works with human expertise rather than replacing it.
This presentation explores how APIs and MARCXML can be used to automate and validate common technical services workflows in a library-system-agnostic way. Using real-world Python scripts developed for catalog maintenance and quality control, the session will demonstrate practical approaches to retrieving bibliographic data via APIs, parsing and validating MARCXML, and programmatically updating records to support tasks such as publisher changes, ceased titles, title changes, and cataloging consistency checks.
Rather than focusing on a specific ILS, the presentation emphasizes transferable patterns: how to think about APIs as workflow entry points, MARCXML as structured data rather than static records, and scripts as repeatable, auditable tools that support human decision-making. The talk will also briefly address how these approaches align with responsible AI use in technical services, including transparency, error detection, and maintaining cataloger authority in the loop.
Attendees will leave with concrete ideas for identifying automation opportunities in their own workflows, regardless of local systems or platforms.
From OCR to MARC: A Governed AI Workflow for Scalable Cataloging Workflows / Alexander Whelan and Zehong Liu
This presentation introduces an in-house, AI-assisted workflow at New York University Libraries which generates brief bibliographic MARC records directly from optical character-recognized image files. In situations where large volumes of library materials remained uncataloged due to staffing constraints or legacy backlogs, cataloging librarians at NYU have critically leveraged AI to support our discovery goals without compromising bibliographic standards or professional judgment. We will share our application of this workflow in two use cases: providing remote support for consortial partners and performing original cataloging for STEM "grey literature."
Our transparent discussion of planning and implementing this project seeks to model a responsible path for integrating AI into cataloging while preserving professional values and control. We will engage our workflow from two complementary perspectives. First, we will share a technical walkthrough with practical insights into batch-creating brief records that can later be reviewed and enhanced per specialized cataloger expertise. Using a university-provided private AI gateway and a governed "CatalogerGPT" prompt, the workflow extracts descriptive elements from photos of a title page and maps them to brief MARC/RDA fields. Outputs are valid records ready for ingestion via standard tools such as MarcEdit. Following the workflow demonstration, we will offer a strategic reflection on the larger organizational impact that a generative AI workflow has on training needs, core competencies, and how staff might engage AI as a support tool rather than replacing human expertise. We show that these practices can continue to support reproducible, auditable, and ethical cataloging work.
Documenting Workflows While in Constant Transition / Rachel Littleton
Workflow documentation is a common and important tool to maintain continuity, yet it can be challenging to maintain while in constant transition. The University of Dayton's Roesch Library recently moved to Ex Libris' Alma LMS as part of a state-wide transition. For over a year, library staff members had to learn a new LMS, make important decisions about their workflow, and document these changes into the shared drive. As these changes occurred, the library also experienced staff turnover as people retired and moved on. As the Library changed, so too did the documentation. The presenter will discuss how their documentation on tasks like ordering, receiving, invoicing, and fund reporting in Alma evolved over this period of time; how they maintained documentation; and what attendees could do in the future with the use of AI and staff collaboration.
------------------------------
Susan Howell
Cataloging and Metadata Librarian
Southern Illinois University Morris Library
------------------------------