Between Automation and Expertise: Employee Sensemaking in Early-Stage Agentic AI Implementation

Abstract

As organizations increasingly invest in AI, understanding the gap between implementation and actual adoption remains a significant challenge. While technical considerations have received considerable attention, employee-level dynamics and the influence of agentic AI remain underexplored in AI adoption research. This thesis examines how employees within a management consultancy firm make sense of the introduction of an agentic AI system during the early stages of organizational implementation, and how this sensemaking process shapes their engagement with the technology. Drawing on sensemaking theory, eight sensemaking mechanisms, and a human-centered AI perspective, a qualitative case study was conducted through six interviews with employees at a Swedish management consultancy firm that recently implemented a locally hosted, agentic AI-powered chatbot. The analysis reveals that employees made sense of the system by drawing on available resources such as familiar generative AI tools, future-oriented beliefs and delegated safety claims, to construct plausible but not necessarily accurate or organizationally aligned interpretations. Employees also turned to colleagues to make sense of the system collectively, which emphasizes the social nature of such sensemaking processes. Furthermore, the specific context of knowledge-intensive work surfaced how employees were not only interpreting a new tool but negotiating their professional identity in relation to a technology that challenges the same capabilities that their professional value is built on. The findings contribute to research on employee-level sensemaking in AI implementation and they offer practical implications for leaders