ArchiMate has recently become the standard language for describing and visualizing Enterprise Architecture models encompassing different domains, their relations, and their dependencies. A few approaches have been proposed for the analysis of ArchiMate models, but they are static and thus focus on structural and static features, leaving aside the analysis of dynamic features. To address this shortfall we propose xArchiMate. It is a model-based approach for executing ArchiMate models and collecting data about dynamic features which is then used, together with static information, to assemble indicators and support decision-making processes. The core of xArchiMate is an execution engine that executes models that conform to the xArchiMate meta-model, which differentiates from the standard ArchiMate meta-model in a few key aspects: it includes behavioral information (e.g. states and interactions) on top of the normal structural information; it includes additional elements that complement and specialize ArchiMate’s base elements; and it only includes elements that have a meaningful behavior and thus are interesting to be simulated. This paper presents xArchiMate, including the meta-model, the execution platform, and a suggested process for creating scenarios and indicators, and then for designing and running experiments to back up decision-making processes on the enterprise.