Graph databases have grown increasingly popular over recent years, as more and more relevant use cases are found that are difficult or even practically impossible to address with relational database management systems (RDBMS). This is because graph databases centre the interconnected nature of real-world data even more so than RDBMS, such that when you care about modelling and querying anything more complicated than a simple, direct relationship, at scale, you had best turn to graph.
The aforementioned use cases include social network analysis, tracking complex medical (often drug) interactions, semantic analysis, and AI/ML (machine learning). The latter is particularly interesting given its recent rise in popularity, both over the last several years as a general trend and the surge that has occurred over the last few months as the general public has gotten access to generative AI technology (such as ChatGPT) for the first time. It is also the primary subject of this paper, in which we intend to examine the interplay that can occur between graph and AI technology, as well as provide some context for the challenges faced in the graph space. We will also highlight a particularly effective example of addressing this interplay, namely Ultipa, a recent addition to the graph scene.