Strategy and Crisis Monitor: An AI-Enabled Autonomous Maritime Monitor for Disruption Intelligence and Operational Resilience

Abstract

Abstract This paper presents the Strategy and Crisis Monitor, an agentic artificial-intelligence (AI) enabled executive dashboard conceived, built, and deployed end-to-end in days to track regional events, incidents, transport disruption, and energy-market signals during a live crisis. The work pursues three objectives: to demonstrate that autonomous AI agents can convert public maritime, energy, and trade data into always-on disruption intelligence; to replace the analyst-curated executive briefing with a live, evidence-linked signal layer operating under human oversight; and to compress executive decision latency from days to minutes. The architecture comprises three operational layers. An ingestion layer of autonomous agents continuously normalizes heterogeneous public sources, including multilateral chokepoint transit data, vessel-tracking feeds derived from the Automatic Identification System (AIS), international oil and gas statistics, and maritime trade and connectivity indicators, supplemented by market context as evidence layers. A classification layer maps incoming data against a rules-based event taxonomy covering chokepoint disruption, export-route stress, gas exposure, port-connectivity deterioration, and regional trade impact. A prioritization layer scores signals using a weighted composite risk index with exponentially weighted moving average (EWMA) and cumulative sum (CUSUM) anomaly detection, and surfaces them through a 60-second executive scan pattern with full drill-down to underlying sources. Deployed live during the March 2026 Strait of Hormuz disruption, when average daily transits collapsed by approximately 95% against the February baseline and the benchmark very large crude carrier (VLCC) rate reached an all-time high, the platform was used to brief senior leadership of a major national maritime and energy logistics operator in near real time. It reduced executive decision latency from a typical 24- to 72-hour analyst cycle to under five minutes per signal, a compression factor of roughly 290 to 860, while integrating more than 40 public indicators. The same detection logic was subsequently applied retrospectively to four historical chokepoint crises, in the Red Sea (2023 to 2024), the Suez Canal (2021), the Strait of Hormuz (2019), and the Panama Canal (2023 to 2024), confirming that the monitored public signals were available weeks to months ahead of official statistics in every case. Those case studies were reconstructed with knowledge of how each episode unfolded and therefore document signal availability and design generality rather than prospective detection performance. The contribution is twofold and is, to the authors’ knowledge, the first implementation of its kind in the maritime-resilience and energy-logistics literature. First, a decision-method contribution: passive public reporting feeds are converted into a structured, always-on resilience instrument fusing maritime, energy, and trade signals in a single operating view. Second, an operating-model contribution: the architecture is public-data-native, lightweight to sustain, and reproducible across operating companies, corridors, and risk domains, and is presented together with the governance, oversight, and control mechanisms required before agentic AI is admitted to a critical operational environment. Both are directly actionable for practicing engineers and planners in operations, strategy, and resilience roles.