Algorithm-based, automated solutions allow designers and engineers to support the tasks of human operators, or transfer parts of their work to a system. However, pushing the limits of automation might make the tasks left for human operators more difficult to accomplish. This thesis zooms in on the work carried out by professional front page editors in a large, online newspaper, to address how an explicit focus on situated human work can contribute to creating alternative automation designs. Concepts and terms from the research field of Computer Supported Cooperative Work have served as a framework for analyzing the tasks and skills of front page editors. I present an analysis demonstrating that the work of front page editors is of a highly judgmental and cooperative nature, and that their decision-making is contingent on several external and changing factors. By contrasting my findings with the design of a recently implemented, semi-automatic curation tool, I highlight several challenges that might arise if we oversimplify human work to accommodate automation. I propose an approach to automation that acknowledges situated human work practices, utilizing ethnographically inspired methods of inquiry akin to those found within the Scandinavian tradition of participatory design. Furthermore, I present a set of prototypes demonstrating that designs for automation can emerge from a situated understanding of front page editing, rather than from the limits of what it is technically possible to automate.