Organizations are increasingly moving toward a team‐based structure for managing complex knowledge in new product development (NPD) projects. Such teams operate in an environment characterized by dynamic project requirements and emergent nonroutine issues, which can undermine their ability to achieve project objectives. Team improvisation—a collective, spontaneous, and creative action for identifying novel solutions to emergent problems—has been identified as a key team‐situated response to unexpected challenges to NPD team effectiveness. Geographic dispersion is increasingly becoming a reality for NPD teams that find themselves needing to improvise solutions to emergent challenges while attempting to leverage the knowledge of team members who are physically distributed across various locations. However, very little is known about how teams' improvisational actions affect performance when such actions are executed in increasingly dispersed teams. To address this gap in the literature, this paper draws on the emerging literature on different forms and degrees of team dispersion to understand how team improvisation affects team performance in such teams. In particular this paper takes into account both the structural and psychological facets of dispersion by considering the physical distance between team members, the configuration of the team across different sites, as well as the team members' perception of being distant from their teammates. Responses from 299 team leaders and team members of 71 NPD projects in the software industry were used to analyze the relationship between team improvisation and team performance, as well as the moderating effect of the three different conceptualizations of team dispersion. Results of the study indicate that team improvisation has a positive influence on project team performance by allowing team members to respond to unexpected challenges through creative and timely action. However, increasing degrees of team member dispersion (both structural and psychological) attenuate this relationship by making it difficult to have timely access to other team members' knowledge and by limiting real‐time interactions that may lead to the development of creative solutions. The results of this research offer guidance to managers about when to balance the desire to leverage expertise to cope with unexpected events. Moreover, the present paper provides directions for future research on improvisation and team dispersion. Future research is encouraged to investigate factors that may help highly dispersed teams to overcome the shortcomings of team dispersion in dealing with emergent events.