Architects encounter the challenge of generating building designs and conceptual customizations during early design stages in an individual and time-efficient manner. Case-Based Reasoning enables support for this process by using similar experienced situations for problem solving of future scenarios. In this context, previously created digital building models can be reused as inspiration and data basis in future projects. This involves comparing the similarities of specific building elements or entire structures within a case database against the design requirements of a future project. This enables an architect to leverage a previously designed building and adapt it to the needs of the current project. However, building models consume substantial storage and assessing similarities often necessitates manual and time-consuming evaluations. Nonetheless, transforming building information models into knowledge graphs based on the international Industry Foundation Classes standard provides low computer memory utilization and fast information processing. For this approach, it is essential to create a case database containing a considerable quantity of data so that problem-specific solutions can be optimally retrieved. This paper presents various approaches for enriching the graph-based case base and examines them for a specific already developed framework. In addition to a manual solution for generating suitable data independently of a digital building model, various Artificial Intelligence methods are assessed and tested to generate synthetic building graphs. This includes an examination of different Generative Graph Models as well as the evaluation of Neural Networks.