THE ROLE OF ARTIFICIAL INTELLIGENCE IN ORGANIZATIONAL DECISION-MAKING: A CONCEPTUAL FRAMEWORK

Abstract

This paper addresses the growing research problem of understanding how and under what conditions Artificial Intelligence (AI) influences organizational decision-making, and aims to develop a conceptual framework explaining how AI capabilities influence organizational decision-making and under what condition these effects occur, that integrates AI capabilities, human actors, and organizational context. Drawing on theories of bounded rationality (Simon, 1957), resource-based view (RBV), and the Technology Acceptance Model (TAM), the framework posits that AI capability (in terms of predictive analytics, machine learning, and decision support systems) are expected to enchance decision speed, decision quality, and strategic alignment in the organizational decision-making. This paper proposed three main propositions: first, that organizational performance is positively related to AI capability via improved decision speed and decision quality (supporting prior empirical findings; Neiroukh, Emeagwali, and Aljuhmani, 2024). Second, that human oversight, transparency, perceived usefulness, and ease of use modulate acceptance of AI in decision settings (drawing from TAM and empirical work; Systems journal, 2025). Third, that ethical, regulatory, and contextual constraints (including AI explainability, bias, compliance) are critical moderators shaping whether AI’s potential is realized as seen in studies of AI governance and decision-making processes (Huang and Niyomsilp, 2025). The framework contributes by synthesizing theoretical perspectives and empirical insights into a model that delineates antecedents, moderators, and outcomes of AI-augmented decision making in organizations. Its significance lies in guiding both researchers and practitioners in diagnosing organizational readiness, designing responsible AI systems, and ensuring that AI serves as augmentation rather than replacement of human judgment. Future empirical validation is suggested, especially in high-stakes and regulated sectors.