AI Agent Operational Lift for Ann Arbor District Library in Ann Arbor, Michigan

Why AI matters at this scale

Ann Arbor District Library (AADL) operates five branches serving a population of over 120,000 in a university-anchored city. With 201–500 employees and an estimated annual budget around $18 million, AADL sits in the mid-market sweet spot where AI can deliver meaningful efficiency gains without the complexity of enterprise-scale systems. Public libraries face growing pressure to modernize digital services while maintaining equitable access, making AI a strategic lever for doing more with constrained public funding.

1. AI-Powered Discovery and Personalization

The highest-ROI opportunity lies in transforming the online catalog. Current keyword-based search often frustrates patrons who can't recall exact titles. Implementing semantic vector search and a recommendation engine—similar to how Netflix suggests content—would dramatically improve the patron experience. This drives digital circulation, increases engagement with underutilized collections, and positions the library as a 21st-century information hub. The technology can be piloted on a subset of the catalog using open-source models, minimizing upfront costs.

2. Operational Efficiency Through Intelligent Automation

Library staff spend significant time answering repetitive questions about hours, event registration, and basic research. A conversational AI chatbot on the website and app can handle 60–70% of these queries instantly, freeing staff for higher-value interactions. Additionally, AI-driven metadata generation for the library's extensive local history archives can reduce cataloging backlogs by auto-tagging images, oral histories, and documents. The ROI here is measured in staff hours reallocated to community programming and patron assistance.

3. Data-Driven Programming and Collection Development

AADL can leverage its rich circulation and event attendance data to predict community needs. Machine learning models can identify emerging interests—such as a spike in gardening or coding topics—and recommend program topics or book purchases before demand peaks. This shifts collection development from reactive to proactive, ensuring budgets are spent on materials that will actually circulate. The impact is both financial (reduced waste on low-turnover items) and mission-driven (better serving community interests).

Deployment Risks Specific to This Size Band

Mid-sized libraries face unique AI adoption risks. First, privacy and ethics: libraries have a professional and legal duty to protect patron reading histories. Any AI system must be designed to avoid creating permanent profiles or sharing data with third parties. Second, digital divide: AI features must not alienate patrons without smartphones or digital literacy; parallel non-digital access must remain. Third, vendor lock-in: many library-specific AI tools are sold as proprietary modules by existing ILS vendors, potentially limiting flexibility and increasing long-term costs. AADL should prioritize open-source or API-first solutions. Finally, staff buy-in: librarians may fear automation threatens their roles. Transparent communication and involving staff in pilot design is critical to positioning AI as an augmentation tool, not a replacement.