The emergence of the Semantic Web in the early 2000s, followed by initiatives like BIBFRAME a decade later, introduced linked data as a transformative approach to knowledge representation and reasoning in libraries. Semantic Web theorists envisioned intelligent agents leveraging structured RDF/OWL to take meaningful actions on behalf of users. However, early implementations based on symbolic logic struggled to fulfill these ambitions. Today, advances in generative AI, particularly transformer based Retrieval Augmented Generation (RAG), introduce the possibility of a new kind of Semantic Web agent, offering enhanced capabilities for discovery and reasoning over structured data. By integrating BIBFRAME-based GraphQL APIs from the Share-VDE discovery system with RAG-powered chat systems, libraries can move beyond traditional search systems to enable deeper, dynamic connections across collections. This presentation explores real-world applications, including Share-VDE’s APIs and experimental integrations of BIBFRAME with Wikidata, demonstrating how libraries can craft LLM-driven prompts to facilitate a paradigm shift from conventional catalog search to intelligent reasoning-based discovery.

This talk will showcase pilot approaches to prompt engineering and retrieval strategies that merge Semantic Web principles with modern AI, providing users with context-aware exploration of knowledge within libraries and across the broader web. Furthermore, by leveraging library-controlled APIs for text inference, institutions can ensure a privacy-conscious search experience, free from the constraints of surveillance capitalism, while fostering deeper engagement with library collections, services, and expertise.