AI Model to Enhance Organizational Decision-Making for Accurate Predictions Using a Machine Learning Algorithm

Abstract

Organizations increasingly rely on data-driven intelligence to navigate complex, time-sensitive decisions. This study develops and evaluates an AI decision-support model based on a Random Forest ensemble tailored to heterogeneous organizational data and operational constraints. The pipeline combines rigorous preprocessing with feature selection and cross-validated training, and is deployed through lightweight, cloud-based APIs and real-time dashboards for seamless integration into existing Decision Support Systems. On a held-out test set, the model achieved 91.2% accuracy, precision = 0.89, recall = 0.91, F1-score = 0.90, and ROC-AUC = 0.94, with an average prediction latency of < 0.5 s per query—suitable for interactive use. Comparative baselines demonstrated materially lower performance: Logistic Regression (accuracy = 85.3%, F1 = 0.83, ROC-AUC = 0.88) and a Single Decision Tree (accuracy = 83.4%, F1 = 0.81, ROC-AUC = 0.86). External validity was examined via three domain-representative simulations. In finance (loan approvals), the model reduced false approvals by 28% versus a rule-based system while maintaining > 90% overall prediction accuracy. In healthcare (30-day readmission risk), it achieved 92% recall, enabling targeted post-discharge interventions and a 17% reduction in avoidable readmissions. In manufacturing (inventory and supply-chain planning), it improved the inventory turnover ratio by 15% and reduced stockouts by 10%, stabilizing production schedules. Across scenarios, automated analytics cut manual assessment time by > 60%, accelerating decision cycles without sacrificing quality. Collectively, results indicate that the proposed ensemble delivers superior predictive power and operational responsiveness relative to conventional models, while remaining adaptable to sector-specific data and workflows. The model’s modular design, fast inference, and integration-ready architecture position it as a practical augmentation to human expertise—enhancing accuracy, timeliness, and consistency of organizational decisions across finance, healthcare, and manufacturing contexts.