Like many organizations, HBS has overlapping data in multiple repositories. These data are maintained in different ways for different business purposes, making it difficult to have a unified, cross-silo view of any given HBS-related entity. The library sought to address this challenge by creating a Proof of Concept for a Knowledge Graph that identifies unique entities (including people, companies and faculty works) across HBS repositories and defines relationships among them. This graph drives the data connections on our website, HBSKnowledge (HBSK). With this integrated structure we have uncovered many of the multiple and varying relationships among data across repositories. We have also set ourselves up to implement inferencing in future releases of the product to further enable the discovery of strategic information at HBS. In addition to structuring our data in graph form, we are using semantic search technology to enhance the search experience. We have leveraged our topic vocabularies, our company authority data, and our faculty & alumni authority data to steer users toward the most relevant information for them. The result is a product that enables discovery of 360 views of entities important to HBS. The HBSK PoC serves as an excellent product for demonstrating the promise of data integration and semantic search, the value of a Knowledge Graph in delivering on that promise, and the talent that exists in libraries for driving content structure and semantic technology. One of the lessons we (re)learned is that any AI initiative is only as good as its data inputs, and a foundation of well-structured, uniquely identified data is essential. To this end, many of our decisions about ontology and vocabulary development were informed by an understanding of library cataloging principles and practices.