ENHANCING DECISION-MAKING IN U.S. ENTERPRISES WITH ARTIFICIAL INTELLIGENCE-DRIVEN BUSINESS INTELLIGENCE MODELS

Abstract

This study addresses a persistent problem in U.S. enterprises: despite heavy investment in AI-augmented business intelligence, many firms still struggle to convert model outputs into faster, higher-quality managerial decisions. The purpose is to quantify how AI-driven BI capability relates to decision quality, decision speed, and near-term performance impact, and to test when those effects are strongest. Drawing on a targeted review of 67 peer-reviewed studies to anchor constructs and measures, we implement a quantitative, cross-sectional, case-based design across 12 U.S. enterprises and 242 decision-making units operating in cloud-enabled enterprise environments. Key variables include an AI–BI Capability Index (feature breadth, workflow integration, automation depth, explain ability in use), organizational conditions (data governance, data quality), a human-capital pathway (user analytics competence), and decision outcomes (quality, speed, performance deltas). The analysis plan pre-registers hierarchical OLS with firm-cluster-robust errors, moderation tests (AI–BI capability by governance and by data quality), exploratory mediation via competence, and robustness checks against multicollinearity, influential cases, and objective-only outcome variants. Headline findings show that higher AI–BI capability is associated with better decision quality and faster signal-to-action cycles, with smaller but meaningful uplifts in performance; effects are significantly larger in units with stronger governance and higher data quality, and are partially transmitted through user competence. Implications are twofold: theoretically, separating capability, conditions, and use clarifies the pathway from AI to decision outcomes; managerially, returns concentrate where governance, data quality, and enablement are treated as first-class complements to modeling and cloud integration.