We are living through the greatest technological transition since industrialisation, with automation precipitating major transformations in the nature and experience of work. For policymakers to craft a fairer future of better work through this transition, they need to understand the different impacts that automation can have, and how these impacts structure different types of risk, in different circumstances, for different groups. In this paper, automation is presented as a process comprising choices made during design, development and deployment of technology. In this regard, while the deployment of technology within firms is most significant, we recognise that aspects of system design which shape impacts are decided before a firm puts them into use. We argue that, rather than outcomes of technology being predetermined, choices made during each of these stages are conditioned by ideas. Ideas shape what we optimise for, and approach to value creation, and in turn specific impacts. While there are risks relating to changing access to or availability of work, linked to the conventionally perceived threat of ‘displacement’ so too can automation affect quality and conditions of work in significant ways, with different outcomes for different groups. By considering automation as a series of decisions which can lead to different types of work transformation, we aim to identify a wider range of intervention points to steer innovation towards beneficial social outcomes, such as a fairer future through better work.