Domestic water consumption projections are needed to manage the urban water supply system in different spatiotemporal resolutions depending on the type of the supported decision. For instance, operational decisions require predictions of fine resolution (hourly or lower), while tactical and strategic decisions are satisfied with predictions of coarser resolutions (monthly or even annual). This work explores the development of the new version of the Urban Water Agents’ Behaviour (UWAB) model, the UWAB 2.0. and its structural changes. These changes are expected to enable the linking of UWAB with machine learning techniques to enhance its projection capacity to support tactical and strategic decisions. UWAB is an agent based modelling tool, which simulates the water demand behaviour of urban households incorporating the effects of climate, water demand management policies and social network to water saving behaviour. This work will present the results of the structural changes of UWAB and the way forward for its development and pilot testing.