Using multi-agent simulation, household preferences, energy use and responses to policies/services will be incorporated into an integrated activity-travel demand model to predict the energy demand at a high level of temporal and spatial resolution. This system will be complemented with a model of distributed energy generation, which represents the possible interactions among local energy suppliers, grid operators and customers. The system design and implementation will be accomplished through co-creation in several demo cities. Experiences will provide the basis to evaluate whether such innovative high-resolution demand predictions provide improved insights into demand flexibility that can be used to balance services in the power grid. This is particularly important in the view of increased penetration of variable renewable generation, electrical mobility and distributed energy generation schemes in cities.