The Impact of Artificial Intelligence on Managerial Decision-Making Effectiveness: Implications for Corporate Governance Practices

Lai, Su Jun (2026) The Impact of Artificial Intelligence on Managerial Decision-Making Effectiveness: Implications for Corporate Governance Practices. Masters thesis, Tunku Abdul Rahman University of Management and Technology.

Text 37 Lai Su Jun.pdf Restricted to Registered users only Download (6MB)

Abstract

Artificial Intelligence (AI) has increasingly transformed managerial decision-making processes by enhancing efficiency, speed, and analytical capability. However, the successful implementation of AI in organisations depends not only on technological adoption but also on the presence of appropriate governance mechanisms that ensure accountability, ethical compliance, and human control. While prior studies have often emphasised the direct benefits of AI adoption, limited research has examined how governance factors influence the effectiveness of AI-driven decision-making, particularly within emerging economy contexts such as Malaysia. This research addresses that gap by investigating the determinants of AI-driven decision-making effectiveness through a governance-oriented framework. Drawing upon Institutional Theory, Dynamic Capability View, and Socio-Technical Systems Theory, this research examines how AI Integration in Decision-Making (AIIDM), Data-Driven Decision-Making Capability (DMC), and Ethical and Security Challenges (ESC) influence the Effectiveness of AI-Driven Decision-Making (EADDM). Human Oversight and Intervention (HOI) and Regulatory and Ethical Compliance (REC) were examined as mediating variables, while Quality of Human Oversight (QHO) and Preventive Regulatory and Ethical Compliance Measures (PRECM) were tested as moderating variables. A quantitative cross-sectional design was adopted, and data were collected from 305 managerial respondents in the Malaysian private sector using self-administered online questionnaires through judgmental and snowball sampling techniques. Data analysis was conducted using IBM SPSS and Hayes’ PROCESS Macro. The findings reveal that governance mechanisms play a more dominant role than technology alone in determining AI-driven decision-making effectiveness. Specifically, REC emerged as the strongest predictor of EADDM, followed by HOI, DMC, and ESC. In contrast, AIIDM did not demonstrate a significant direct effect on effectiveness. Mediation analysis confirmed that both HOI and REC significantly mediate the relationships between AI-related factors and EADDM, indicating that governance mechanisms serve as key transmission pathways through which AI capabilities generate value. Moderation analysis further revealed that higher levels of QHO and PRECM buffer the dependence of effectiveness on oversight intensity and reactive compliance, respectively. This research contributes theoretically by advancing a governance-driven perspective of AI effectiveness and integrating three complementary theories into a unified framework. Empirically, it provides evidence from Malaysia as an emerging economy context. Methodologically, it demonstrates the usefulness of combining regression, mediation, and moderation analyses in AI governance research. Practically, the findings suggest that organisations should prioritise governance readiness, ethical compliance, proactive controls, and meaningful human oversight when implementing AI systems. The research concludes by acknowledging its limitations and recommending future studies involving longitudinal designs, objective performance measures, and more advanced modelling approaches.