Spatial complexity, defined by a multitude of interconnected variables, is recognized as one of the most challenging forms to predict. Given the inherent unpredictability of the future and the limited utility of forecasting models for long-term predictions, this paper proposes an integrated approach for navigating various potential futures. For the development of spatial scenarios, various methods are currently adopted, including the Real-Time Spatial Delphi approach. This method involves leveraging expert judgments through a real-time process to attain a spatial consensus on the territory. Nevertheless, the outputs of this method encompass spatial analyses and statistical indicator results which, from a communicative standpoint, may fail to capture the attention of non-experts, such as citizens or policymakers. To overcome this challenge, the paper suggests integrating Generative Adversarial Networks models to produce realistic visualizations of scenario-based policies. Through a case study conducted in Massa, Italy, as part of the EU H2020 SCORE project, the effectiveness of this approach is demonstrated. The findings highlight the significance of incorporating artificial intelligence methods to enhance the communicative aspect of spatial scenarios, providing efficient visual representations of scenarios and emerging policies ready for assessment.