AI Becoming Practice : Professional Work and Autonomy in Technology Consulting

Sammanfattning

This thesis examines how AI-augmented work reshapes professional practices and perceptions of autonomy among tech consultants. Drawing on sociomaterial perspectives, it conceptualizes AI as constitutively entangled in professional practice rather than as an external tool. The empirical material consists of 21 semi-structured interviews with tech consultants in Sweden, complemented by document analysis of AI tool documentation and interfaces. Three orientations toward AI were identified in current practice, instrumental, relational, and embodied, differing not only in degree of integration but in where agency resides and what professional autonomy consists of in relation to AI. A fourth anticipatory orientation captures how consultants actively reconfigure their practices in response to expected institutional consequences of AI adoption. The thesis makes three contributions. It extends Scott and Orlikowski’s (2025) versioning concept by showing that enactment fluidity operates within individual practitioners, not only across a population. It introduces counter-mimesis to describe how knowledge workers respond to automation threat through deliberate professional distinction rather than algorithmic conformity. It extends Orlikowski and Scott’s (2023) digital undertow framework by showing that the undertow can become visible to practitioners, enabling proactive adaptation. These latter two contributions form an integrated finding: consultants who perceive the undertow are precisely those who engage in counter-mimesis, with domain expertise, in their own accounts, as the condition making that perception actionable.