Lucy Campbell and Keven Jeffery write: “The predominance of artificial intelligence (AI) tools has resulted in challenges for library and information professionals, particularly in academic research libraries. For example, these tools frequently hide the underlying search logic from the user, creating an opacity that misses opportunities for learning about advanced search. Results are siloed to specific platforms instead of surfacing all content owned or subscribed to by a library. This case study outlines a project at San Diego State University Library to build an AI discovery assistant that can surface search logic, retrieve results from across holdings, and provide learning opportunities for users.”