Action research (AR), a methodology based on active collaboration between researchers and practitioners, and aiming at systems change, appears to be an appropriate approach to respond to the increasing demand for impact-driven studies. Yet, such intent to produce situated impact also comes with two significant limitations. First, because of the presupposition that collaborative, problem-solving orientated research fails to produce generalizable theoretical contributions, action research studies rarely achieve publication in the most influential academic journal. Second, the tacit assumption that its impact will inherently be positive impact, minimizing the possibility of negative consequences of the intervention. Drawing from the empirical experience gained in an action research study, we address both issues, showing that – since action research's deep engagement with organizational challenges surfaces organizational paradoxes – persistent, contradictory, and interrelated tensions inherent to organizing, both its heuristic potential and the risks that its implementation should not be underestimated for two interdependent reasons. On the one hand, paradoxes can be instrumental to question and extend existing theories and to uncover uniquely rich data. On the other hand, researchers must be aware that poor management of paradoxes in research interventions can lead to adverse outcomes, including unexpected conflicts and resistance. Our empirical findings offer new insights into the role of paradox management as a fundamental methodological practice to foster responsible action research practices.
