Redefining Middle Management: How Generative AI Reshapes Roles and Competencies

This ongoing research investigates how Generative Artificial Intelligence (GAI) technologies are reshaping the roles and competencies of middle managers. While much of the existing literature emphasizes macro-level impacts of AI—such as productivity gains and strategic transformation—this study focuses on micro-level changes, particularly managerial adaptation, decision-making, and skill reconfiguration. Grounded in activity theory (Engeström, 1987, 2001) and informed by a critical realist epistemology, the study adopts a qualitative, inductive approach. Data is being collected through semi-structured interviews with middle managers from three contrasting organizational contexts: a large enterprise in the energy sector (30 interviews), a medium-sized digital services firm (15 interviews), and a small telecommunications company (15 interviews). This multi-site design enables the analysis of systemic tensions across different governance models and organizational cultures. Preliminary findings reveal that middle managers are often at the forefront of GAI experimentation, initiating Bottom-Up innovation processes outside official IT channels. These informal practices—frequently associated with Shadow IT—allow for agile problem-solving but also pose significant risks in terms of data security and strategic misalignment. Managers report that GAI tools not only automate routine tasks but also support cognitive structuring, thereby transforming how they plan, communicate, and lead. The integration of GAI calls for hybrid competencies that go beyond technical know-how. Conceptual skills (e.g., sense-making and system thinking), human skills (e.g., empathy, leadership, conflict mediation), and ethical awareness are increasingly vital. Middle managers thus emerge as key facilitators of digital transformation, mediating between frontline innovation and top-down strategic frameworks. This paper contributes to the literature by offering a multi-dimensional reading of managerial transformation through the lens of activity theory. It proposes actionable insights for organizations aiming to responsibly integrate GAI, emphasizing the need for adaptive governance, targeted upskilling, and the creation of experimental safe zones. Future work will extend the dataset and explore longitudinal trajectories of GAI adoption and governance.