The Organisational Challenge of AI: Implementation in Swedish Multinational Service Firms

Abstract

Artificial intelligence is increasingly central to value creation in multinational service firms. Yet its implementation remains uneven across firms with comparable access to AI technology. While existing research links AI to productivity gains, most of it is grounded in manufacturing contexts. This leaves a gap in understanding how service-based MNEs embed AI into routine work across international operations. This thesis examines how Swedish multinational service firms initiate, govern, and manage AI implementation. It combines secondary data from Statistics Sweden with semi-structured interviews across five firms. The study draws on value configurations, the resource-based view, absorptive capacity, and dynamic capabilities. The central finding is that AI implementation is fundamentally an organisational challenge rather than a technical one. Access to AI tools was not the binding constraint in any of the five cases. What differentiated firms was their internal capacity to convert awareness of AI's potential into embedded operational practice. This capacity is governed by a chain of organisational conditions: the firm's value configuration, its internal resource base, its governance architecture, and its dynamic capabilities. These conditions must function simultaneously across multiple national contexts, making AI implementation in multinational service firms a distinctly organisational and cross-border challenge.