The European welfare state undergoes a significant transformation. On one hand, the rise of the neoliberalism prompted austerity and more means-testing of claimants. On the other hand, new digital technologies for collecting, sorting and processing (big) data allow welfare authorities unprecedented reach into the private lives of claimants, recasting suspicion on citizens through an automated, algorithmic decision-making process that determines their eligibility to the most basic of needs. This paper examines the automatization of welfare provision and the subsequent impact on the discretion of social workers. Drawing on ethnographic research at Centres for Social Work in Slovenia (Europoe), this paper argues with and against Buffat’s (2013) ‘curtailment thesis’, which posits that digitalisation causes a reduction in frontline policy discretion. It highlights the variety of ways in which automatization of welfare provision reduces social workers’ discretion, but also creates new ‘sites of discretion’ (McCann 2023) beyond the scope of the algorithm or that of legislative decrees. It then explores the significance of these transformations on how the digital welfare state is perceived by both claimants and social workers. Lastly, it presents a number of guidelines that can preserve the social nature of welfare institutions and maintain their commitment to the citizens they serve.