Implementation and Real-World Evaluation of an AI-Based Decision Support Model for Data-Driven Business and Infrastructure Systems

Abstract

This study addressed the practical problem that many organizations invest in artificial intelligence driven decision support systems, yet evidence of their real-world effectiveness across business and infrastructure settings remains limited and often fragmented. The purpose of the research was to evaluate how an AI-based decision support model contributes to decision quality, operational efficiency, and organizational performance in data-driven environments through a quantitative, cross-sectional, case-based design. The study used a structured five-point Likert scale questionnaire administered to 210 respondents drawn from cloud-enabled and enterprise-oriented organizational cases, including private sector enterprises, public sector organizations, and hybrid institutions, with 108 cases from business systems and 102 from infrastructure systems. The key variables examined were AI Model Capability, Implementation Readiness, Trust in AI Recommendations, Perceived Usefulness, System Reliability, Decision Quality, Operational Efficiency, and Organizational Performance. The analysis plan combined descriptive statistics, correlation analysis, and multiple regression modeling. The findings showed that respondents rated the model positively across all major constructs, with Decision Quality recording the highest mean (M = 4.25, SD = 0.58), followed by Perceived Usefulness (M = 4.22, SD = 0.57), AI Model Capability (M = 4.18, SD = 0.61), Operational Efficiency (M = 4.11, SD = 0.64), and Organizational Performance (M = 4.07, SD = 0.66). Correlation results indicated strong positive relationships, particularly between Decision Quality and Organizational Performance (r = .73), Perceived Usefulness and Organizational Performance (r = .71), and Decision Quality and Operational Efficiency (r = .69). Regression findings further revealed that AI Model Capability significantly predicted Decision Quality (β = .31, p < .001), Implementation Readiness significantly predicted Operational Efficiency (β = .29, p = .001), and Decision Quality significantly predicted Organizational Performance (β = .34, p < .001). The study implies that AI decision support creates the greatest value when technical capability is supported by organizational readiness, reliable system performance, and human oversight, making AI most effective as an augmentation tool rather than a replacement for human judgment.