Reconfiguring Managerial Problem-solving: The Role of Artificial Intelligence

Sammanfattning

The increasing integration of artificial intelligence (AI) into managerial work is transforming how organizations approach problem-solving and decision-making. While prior research has primarily examined how AI supports individual decision-making stages, limited attention has been given to how AI influences the overall structure and dynamics of managerial problem- solving processes in practice. This study addresses this gap by examining the research question: How does artificial intelligence reconfigure the problem-solving of middle managers? Drawing on ten semi-structured interviews with middle managers across different industries, this qualitative study applies an abductive approach and Kathy Charmaz’s constructivist grounded theory methodology. The analysis is guided by Simon’s Intelligence–Design–Choice (IDC) model to examine how AI is integrated across different stages of managerial problem- solving. The study identifies two main dimensions: the reconfiguration of problem-solving processes and the mechanisms shaping AI use. The findings show that AI contributes to three main reconfiguration changes. First, managerial problem-solving becomes increasingly iterative, characterized by repeated human–AI interaction loops involving prompting, evaluation, and refinement. Second, the early phases of problem-solving are expanded, enabling faster problem understanding, broader exploration of alternatives, and enhanced idea generation. Third, the overall process becomes more non-linear, with flexible movement across stages and increasingly fluid boundaries between them. In addition, the study identifies four key mechanisms influencing AI use: trust calibration, conditional deployment, organizational constraints, and human decision authority. The findings further identify Human–AI Co-Development as a central interactional mechanism through which managers and AI systems jointly shape problem-solving activities throughout the process.