AI-POWERED BI DASHBOARDS IN OPERATIONS: A COMPARATIVE ANALYSIS FOR REAL-TIME DECISION SUPPORT

Abstract

This study addresses a practical problem in operations: decision latency and variable decision quality when teams rely on visually dense dashboards under high data volume, velocity, and variability. The purpose is to quantify how AI-integrated business intelligence dashboards support real-time decision making across organizations. Using a quantitative, cross-sectional, case-based design, we analyze six production cloud and enterprise cases in manufacturing, logistics, healthcare operations, tech-enabled services, retail fulfillment, and utilities, with 168 active users as respondents. Key variables include an AI Integration Index (forecasting, anomaly detection, prescriptive recommendations, natural-language interaction, explainability), user perceptions (perceived usefulness, interpretability, trust, workload), outcomes (decision latency, decision accuracy or confidence), and contextual controls (data quality, dashboard tenure, organization size, training, analytics proficiency). The analysis plan combines descriptive statistics and correlations with multivariate regression using HC3-robust errors and case fixed effects, mediation tests via bootstrap for perceived usefulness, and moderation tests for data quality. Headline findings show that higher AI integration is associated with materially faster decisions and higher confidence, with perceived usefulness transmitting much of the effect on confidence and dependable data quality strengthening the speed benefits; interface workload relates to slower action. Implications are concrete for architects and managers: prioritize pipeline timeliness and semantic clarity, expose compact on-demand explanations and uncertainty cues, control alert and visual clutter, and connect predictions to guarded prescriptive actions. The literature review synthesizes 57 peer-reviewed papers to ground constructs, measures, and mechanisms used in this comparative evaluation.