Managerial Use of Generative - AI Enablers and Constraints in a Consulting Context

Abstract

Generative AI tools like ChatGPT can transform how managers make decisions, yet adoption at the managerial level remains limited. Why? Through interviews with nine managers at Rejlers and validation from a generative AI transformation specialist, we identified three key constraining areas. Trust develops through domain expertise, creating a paradox where knowledge enables verification of generative AI outputs while simultaneously making managers more critical of them. Managers with stronger expertise feel confident using AI because they can evaluate accuracy, yet this heightens their scrutiny of AI-generated content. Boundaries are established based on contextual understanding. Managers readily delegate routine tasks to generative AI while reserving complex, relationship-sensitive decisions for human judgment. AI's lack of organization-specific context limits its utility for company specific challenges. Practical barriers impede adoption, including insufficient integration with existing systems, established work habits, and limited awareness of generative AI capabilities. These factors often prevent adoption even when managers recognize potential benefits. From these insights, we developed the Managerial AI Interaction Framework illustrating how managers progress through four stages: awareness and experimentation, competence building, workflow integration, and boundary-aware collaboration. Our research suggests that effective human-AI collaboration requires addressing trust dynamics, establishing appropriate task boundaries, and overcoming practical integration barriers rather than focusing solely on technical capabilities.