In the Nordic countries, which are known for their expansive welfare systems, public administration is increasingly being viewed as a potential automation site. Data-driven process automation is believed to alleviate administrative drudgery and support a goal-driven, efficient public sector. In the implementation of automation systems, a variety of stakeholders participate, including corporate actors, managers, consultants, politicians and civil servants. As stakeholders, these groups have diverse perspectives and expectations regarding the anticipated future of automation and its role in the organization of public services. The idea of automated work processes is accepted by some as a way to boost efficiency, productivity, and precision through algorithmic data processing; others, however, see such systems as ways to let professionals spend less time on repetitive, rule-based, and seemingly tedious tasks so that they can focus on their core professional practice. This paper engages with the tensions between the different ways of imagining automatized futures by reporting from ethnographic research with stakeholders from roughly ten municipalities in Sweden, as well as digital technographic studies of two world- leading work automation platforms. Using the sociology of expectations as a theoretical framework to illuminate these tensions, this paper explores the role of anticipated automation when organizations imagine change. The paper shows that automation can be used to help envision, think about, and communicate organizational change without involving any technology, but rather as a metaphor to which future-making practices can be attached and legitimized. Work automation is understood in the paper as fundamentally social and structured as well as affected by expectations regarding future benefits and potential risks. By explicitly paying attention to expectations and experiences, the paper allows for an interrogation of emerging automated work futures that goes beyond techno- optimism and economic-political goals of efficiency and optimisation, not the least by illustrating work automation’s situated and contingent nature.