In the Architecture, Engineering, Construction, and Operations (AECO) industry, digital models are crucial throughout a building's lifecycle, particularly in the initial design phases for three-dimensional visualization, complex analysis, simulation, and virtual exploration. Addressing the complexities of net-zero strategies and sustainability requires extensive investigations and storage-intensive digital models. This paper explores advanced methods, such as Generative Artificial Intelligence (GenAI), to enhance energy-efficient digital designs. We propose a memory-efficient approach to convert building models to Industry Foundation Class (IFC) format using knowledge graphs, facilitating quick retrieval and display of detailed information to identify inefficient building components. These components can be replaced with sustainable alternatives from similar projects, though individual modeling of alternatives is often necessary due to limited data availability in planning offices. Our study aims to populate a knowledge base with synthetic graphs for decision support, generating synthetic data aligned with the IFC hierarchy to aid in evaluating alternative designs. Using GenAI methods, we generate and test synthetic data within an IFC compliant graph, demonstrating the evaluation of building designs through a case study on achieving diverse targets. This combination of GenAI and data management techniques aims to expedite the design process, meeting diverse sustainability goals more efficiently.