The Role of Machine Learning Model Interpretability in Improving the Quality of Strategic Decision-Making

Abstract

This study aims to explore the role of machine learning model interpretability in improving the quality of strategic decision-making. The study addresses the growing use of machine learning in organizational decision processes and the problem of limited transparency in AI-generated recommendations. A qualitative case study approach was used to examine how decision-makers understand, evaluate, and apply model explanations in strategic contexts. Data were collected through semi-structured interviews, non-participant observation, and document analysis involving managers, strategic planners, data analysts, data scientists, AI project leaders, risk officers, and policy decision-makers who had direct experience with machine learning-based decision support systems. Thematic analysis identified four main themes: interpretability as a basis for trust formation, interpretability as a bridge between technical analysis and managerial judgment, interpretability as a tool for reducing uncertainty, and interpretability as a mechanism for accountability and justification. The findings show that decision-makers use model explanations not merely to understand algorithmic outputs, but also to negotiate meaning, validate recommendations, and align AI-supported insights with organizational priorities. This study contributes to human-centered explainable AI and strategic decision-making theory by showing that interpretability functions as a social, managerial, and organizational process. The findings imply that organizations should develop AI systems and decision policies that prioritize clear explanations, contextual relevance, uncertainty disclosure, and accountability. Future research should compare interpretability practices across sectors and examine how data literacy and organizational culture influence AI-supported strategic decisions.