There seems to be a need to improve innovation strategies in sustainable architecture due to the high carbon emissions caused by the architecture sector which amounts to 40% of global carbon emissions. Despite the efforts towards sustainability and reducing emissions, a major challenge is the building performance gap (prediction, expectation, and outcome gap) between design intentions and the real-world construction outcome. This research focuses on the potential of artificial intelligence when it intervenes in the early design and decision-making stages, specifically the generative design platform Hypar to investigate the potential of mitigating performance gaps in sustainable architecture design. The study demonstrates the application of Hypar to a case study to measure its potential in improving design decisions when incorporating real-time environmental data to enhance performance efficiency and sustainability. The results indicate that AI potential can enhance design accuracy when predicting actual outcomes early, indicating the possibility of reducing the need for post-construction modifications. The research also demonstrates the transformative potential of AI in reducing performance gaps and improving building performance. Also, AI seems to have the capabilities to contribute to improving the future of sustainable architecture when architects are provided with generative design tools and reliable data.