Algorithms shape our private and professional lives. They are woven into the very core of our society. The purpose of this research is to elucidate how routines and everyday work practices of organizations are shaped and transformed by the increasing use of learning algorithms. Building on post-structuralist approaches, the project examines the agentic and temporal aspects in the unfolding of human-machine interactions. Combining ethnographic and digital trace data, it follows action patterns at a start-up and an established firm, studying how "algorithmic routines" emerge, develop, collapse. Particular attention is given to the performativity of humans and machines as well as their rhythmic entanglement. The findings will provide insights that allow organizations and policy makers to better understand the role algorithms play in organizing by moving beyond conceptualizing them as tools that need to be managed.