AI-Enabled Decision Intelligence for Information Systems Management: A Predictive Framework for Organizational Performance and Resource Optimization

AI is playing more and more of a part in the management of information systems, enhancing predictions, operational decisions and resource allocation. The ServiceNow incident-management event log used for the quantitative secondary-data approach had 141,712 records from 24,918 incidents. Event records were combined at the incident level and operational variables associated with priority, urgency, assignment and reassignment, reopening, category and service outcomes were made ready for analysis. Linear, ensemble and gradient-boosting models were built to predict the time to incident resolution and risk of breaching SLAs. Random Forest had an R² of 0.8318 when predicting resolution-time and a mean absolute error of 60.67 hours, while tuned XGBoost had an F1 score of 0.9571 when classifying SLA. The factors that were identified as important drivers of performance were reassignment frequency and system modifications, characteristics of the assignment group, priority, category, temporal factors, and reopening behaviour, as determined by explainability analysis. However, incidents that were repeated with reassignment had longer resolution times and higher SLA breaches, suggesting clear opportunities for workload balancing, and early intervention during incident handling. The resulting framework integrated predictive analytics, explainable AI, and resource prioritization to convert operational data into actionable decision intelligence. The results highlight the effectiveness of an AI-backed information system management solution in enhancing the effectiveness of services, managing incidents, and prioritizing technical resources. The framework also offers a foundation for scaling predictive decision intelligence across the organization and to real-time resource-management contexts.