Proceedings of the International scientific and practical conference ―Modern Science: Challenges and Perspectives‖ (February 9-11, 2026) / Publisher website: www.naukainfo.com. - London, United Kingdom, 2026. - 121 p.

8 Approaches to diagnostics and foresight under turbulence Methodologically, modern enterprise diagnostics increasingly combines performance measurement with causal and process-oriented interpretation: KPI systems and financial ratios remain necessary but insufficient unless complemented by process mining, anomaly detection, constraint identification, and risk sensing that explains ―why‖ deviations occur and ―where‖ corrective leverage exists. Foresight, in turn, shifts from single-line forecasting toward multi-future reasoning, where scenarios and weak signals are used not only to anticipate outcomes but to design strategic options and ―decision triggers‖ that define when the organization should scale, pivot, pause, or terminate initiatives. Under turbulence, the value of foresight lies less in predictive accuracy and more in preparedness, i.e., the ability to map uncertainty into structured choices and to pre-commit governance responses to observable indicators. This implies that diagnostics and foresight should be co- designed: diagnostics produces reliable signals and explanatory narratives, while foresight transforms them into options, contingencies, and thresholds that guide action without requiring full certainty. Digital infrastructures as an enabler of continuous diagnostics and foresight The shift toward continuous diagnostics and foresight requires architectures that reduce latency between events and interpretation and that sustain analytical quality over time. Event-streaming patterns, including Kafka-based pipelines, support near- real-time ingestion of operational, customer, and external signals and allow multiple analytical services to consume the same event streams for monitoring, alerting, and model inference. Cloud-native data foundations, implemented on platforms such as Amazon Web Services, Microsoft Azure, and Google Cloud, make it feasible to combine scalable storage, elastic processing, and managed analytics and machine- learning services, thereby supporting both exploratory diagnostics (data-lake logic) and standardized reporting (data-warehouse logic). In turbulent environments, however, the decisive factor becomes lifecycle reliability: models and indicators drift as markets, behaviors, and processes change; therefore, MLOps discipline (monitoring, drift detection, retraining rules, lineage, and governance) is not optional