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.

6 signals; scenario analysis; explainable AI; process mining; MLOps; ForesightOps. Introduction Global crises and technological turbulence shorten the ―half-life‖ of strategic assumptions and increase the costs of delayed detection of deviations. In such contexts, enterprises cannot rely on episodic diagnostics or periodic forecasting as isolated analytical exercises; they require continuous diagnostics that explains why performance deviates and where constraints are emerging, and systematic foresight that explores alternative futures and converts uncertainty into structured strategic options. This shift moves diagnostics and foresight from the periphery of strategy work toward the core of enterprise development governance, where data pipelines, model lifecycle management, and transparency requirements become decisive for the quality and speed of decisions. Accordingly, the research task is not only to describe methods of diagnostics and foresight, but to explain how they can be integrated into a maintainable operating model that links sensing, interpretation, decision-making, and learning. Literature review In work [1], Fraunhofer Institute for Technological Trend Analysis (Fraunhofer INT) presents data-driven foresight as a scalable approach that combines large heterogeneous datasets with analytical and expert workflows (including the KATI system) to support technology and innovation scanning; however, the gap is that such platforms are often discussed as powerful analytical environments without a sufficiently explicit mechanism that operationalizes ―foresight outputs‖ into enterprise development decision rights, thresholds, and recurring governance cadence. In work [2], Marinković et al. consolidate corporate foresight research into an integrative view of antecedents, tools, moderators, and outcomes, but foresight is still frequently positioned as a parallel function, and the unresolved gap remains the translation layer from scenarios and weak signals into disciplined strategic choices such as option portfolios, trigger-based escalation rules, and measurable renewal routines. In work [3], Nannini et al. synthesize the rapidly growing intersection of explainable AI and process mining and show that interpretability is becoming a