Organizational routines are ubiquitous stores of knowledge in organizations. Although routines enable consistent performance on tasks over time, routines might hinder adaptability by promoting inertia and rigidity. In this dissertation, I develop how routines could facilitate adaptation in organizations and foster successful performance on novel tasks. I argue that teams are changed in the process of using routines. I develop and test theory arguing that routines can facilitate the development of transactive memory systems (TMS), collective systems for encoding, storing, and sharing knowledge. I propose that routines provide a structure within which team members can learn about one another’s skills. Thus, routines can build a team’s TMS, which can improve performance on novel tasks. I use a mixed-methods design to both develop and to test theory. In Study 1, I performed 30 semi-structured interviews with United States Marine Corps officers. Based on interview findings, I hypothesize that teams that use routines will develop stronger transactive memory systems than teams that do not. Consequently, I hypothesize that teams that use routines will perform better on novel tasks due to the TMS they have developed. In Study 2, I developed a laboratory study in the cybersecurity context to test the hypotheses. Two hundred and thirteen participants in 71 teams were randomly assigned to perform a task with a routine or without a routine, and then perform a novel task. Results provide evidence to support the hypotheses.