ARTIFICIAL INTELLIGENCE–ENHANCED CYBERSECURITY FRAMEWORKS FOR REAL-TIME THREAT DETECTION IN CLOUD AND ENTERPRISE

Real-time threat detection in cloud and enterprise settings remains constrained by high alert noise, data drift, and limited analyst capacity; organizations need evidence on whether deeper AI integration improves detection timeliness and operational quality. This study’s purpose is to quantify the association between AI-enhanced cybersecurity frameworks and measurable outcomes under routine operations. We adopt a quantitative, cross-sectional, case-based design spanning 18 heterogeneous deployments across cloud-first, hybrid, and on-premises environments. The sample consists of cloud and enterprise cases that meet inclusion criteria for active SOCs, centralized logging, and at least one AI-driven detection or orchestration component. Following a targeted review of 100 peer-reviewed papers to ground constructs and measures, we operationalize an AI Integration Index and model its relationship to key variables detection latency, precision, recall, F1, false-positive rate, and mean time to respond using a pre-registered analysis plan: descriptive profiling, correlation screening with multiple-comparison control, robust OLS and median quantile regression for continuous outcomes, beta or fractional logit models for rates, interaction tests with cloud maturity, influence diagnostics, leave-one-case-out validation, and bootstrap intervals. Headline findings show that higher integration aligns with materially shorter detection latency and MTTR, higher precision and F1, and lower false-positive rates, with effects strongest in mature cloud contexts where control-plane observability and API-mediated enforcement are pervasive. Implications for practice are to treat AI as a system capability spanning fusion across telemetry, model freshness with analyst feedback, drift monitoring, and graded, reversible SOAR automation that links detection confidence to proportionate action, thereby reducing dwell time without sacrificing oversight.