The Evolution of Organizational Decision-Making : Exploring How Generative AI Can Enhance Decision-Making in Multinational Corporations

Sammanfattning

In the era of Big Data, organizational decision-making is becoming increasingly challenging due to information overload. This becomes further complicated in multinational corporations (MNCs), where decentralized structures, siloed information, and cross-border operations are common. While generative AI offers new opportunities to support decision-making, its role in such complex environments is not yet fully explored. This thesis investigates how generative AI can enhance decision-making in MNCs. Drawing on the People, Process, Technology model, the study explores generative AI in decision-making from a sociotechnical perspective. Through a qualitative approach combining traditional and AI- conducted interviews with managers and experts, it identifies key obstacles in the organizational decision-making and examines how generative AI can be applied to enhance these. Findings show that decision-making in MNCs is often constrained by fragmented structures, siloed workflows, and limited access to relevant data, factors that reinforce bounded rationality. Generative AI can address these challenges by improving automation and augmenting complex decisions, marking a shift from bounded rationality to bounded trust where users hesitate to rely on AI despite its technical accuracy. However, realizing the potential of generative AI depends on organizational readiness, data governance, and systems designed for trust, transparency, and contextual relevance. To succeed, generative AI must be adapted to decision type and risk level, supported by context-specific data, aligned with existing organizational structures, and designed to balance user acceptance. These insights position generative AI as a promising but context- sensitive tool for decision support in MNCs, offering both practical and theoretical guidance for its implementation in complex organizations. Future research should further explore the concept of bounded trust in human-AI decisions- making, and the role of synthetic data in decisions where real data is limited. It also highlights the need to examine how regulatory frameworks affect implementations of AI in MNCs decision- making processes, and further investigation into the use of AI interview agents as scalable tools for qualitative data collection.