From snapshot to signal: How maritime intelligence is transforming underwriting risk models

Marine underwriting has always been a discipline shaped by experience and precedent. In a rapidly shifting risk landscape, that foundation is being tested in new ways.

For decades, the underwriting process for hull and machinery, P&I, war risk, and cargo lines has followed a broadly consistent methodology: gather vessel particulars, review loss history, consider flag state and classification society, assess trading area, and price accordingly. A risk is evaluated at a point in time. A policy is bound. The underwriter moves on.

That approach worked reasonably well in a slower, more transparent shipping world. It is increasingly inadequate in the one we operate in today.

‍

The limits of the static risk model

Marine underwriting has traditionally relied on what might be called a "snapshot" assessment — a cross-sectional view of a vessel or fleet captured at inception or renewal. The problem is that ships move. Operators change. Trading patterns shift. Cargo origins are obscured. And between annual renewals, a vessel's actual risk profile can change dramatically.

Global marine insurance premiums reached a record $39.9 billion in 2024, yet growth slowed to just 1.5% — down from 5.9% the prior year and 8.3% the year before that¹. That deceleration, combined with intensifying competitive pressure, means underwriters cannot afford to carry mispriced risk into their books. The margin for error is narrowing.

Meanwhile, the risk environment is becoming structurally more complex. War risk corridors, deceptive shipping practices, shadow fleet expansion, GNSS manipulation, and geopolitical flux are now permanent features of the underwriting landscape — not episodic disruptions. These are not risks that a static annual snapshot can adequately capture.

‍

What behavioural-based risk modelling means in practice

Behavioural risk modelling shifts the underwriting lense from what a vessel is to how it operates. Rather than relying solely on vessel specifications, historical loss ratios, and flag state classification, it integrates continuous operational signals — patterns of movement, port call sequencing, AIS transmission behaviour, ownership network changes, cargo type and origin, and deviation from established trade routes.

The key distinction is temporal. A snapshot tells you where a vessel was and what it looked like at a given moment. Behavioural modelling tells you whether its conduct over time is consistent with the risk you thought you were insuring.

The International Union of Marine Insurance has noted that modern risk assessment is shifting toward dynamic matrices that evaluate exposure across multiple dimensions: route volatility, port congestion, geopolitical instability, and weather anomalies — enabling underwriters to move beyond historical loss ratios toward predictive, scenario-based pricing².

This is not a theoretical evolution. It is already happening in the market.

‍

The signal set underwriters should be monitoring

AIS behaviour as a risk indicator

Automatic Identification System data has historically been used as a navigation and tracking tool. Its value to underwriters has been under-appreciated. A vessel's AIS transmission pattern — when it goes dark, where, for how long, and in what sequence — is among the most informative behavioural signals available.

AIS manipulation has become endemic in certain trade segments. According to Kpler analysis, dark STS transfers surged sharply over the past year, overtaking AIS spoofing as the primary indicator of deceptive activity — a direct reflection of an expanding shadow fleet and escalating efforts to conceal sanctioned cargo movements. The Gulf of Oman has emerged as a particular hub for ship-to-ship transfers between sanctioned and non-sanctioned tankers, with increasingly sophisticated tactics used to disguise transshipment activity.

For underwriters, a vessel that repeatedly goes dark in high-risk zones — the Laconian Gulf, waters off Malaysia and Singapore, West African anchorage areas — is not exhibiting a technical anomaly. It is exhibiting a behavioural pattern that should inform both acceptance decisions and policy conditions.

Shadow fleet exposure and associative risk

The shadow fleet represents one of the most significant structural shifts in maritime risk of the past decade. Since Russia's full-scale invasion of Ukraine, shadow fleet tankers have grown to represent approximately 17% of the global tanker fleet — a doubling since 2022. By mid-2025, more than 400 oil tankers were estimated to be operating under opaque ownership without Western insurance.

The underwriting implications extend beyond vessels that are obviously non-compliant. Associative risk — exposure arising from proximity to, or operational overlap with, sanctioned or grey-zone vessels — is a growing concern. Ships operating in fleets alongside shadow fleet tankers, employing management firms that service those networks, or regularly calling at transshipment hubs known for sanctions evasion, carry elevated exposure even when their own compliance record appears clean.

The OFAC 2025 advisory formalised this concept, requiring maritime stakeholders to research all vessels involved in successive ship-to-ship transfers — not only by name, but using multiple identification and location data points. For underwriters, this is a direct signal that vessel-level due diligence is no longer sufficient without network-level analysis.

Ownership opacity and beneficial ownership changes

One of the clearest predictors of elevated risk is a change in beneficial ownership structure — particularly where that change involves a shift to opaque jurisdictions, shell company layers, or rapid reflagging. Shadow fleet operators routinely replace sanctioned vessels by purchasing tonnage from secondary markets and reregistering them under obscure flag states. Vessels in the EU's 14th sanctions package were frequently identified on the basis of irregular and high-risk shipping practices, not simply cargo origin.

An underwriting team that monitors post-inception ownership changes and flag state transitions has a materially more accurate view of its exposure than one that relies solely on inception-date particulars. This is basic risk management that maritime intelligence data now makes operationally feasible.

Port call sequencing and geographic risk profiling

A vessel's trading pattern tells a story. Sequential calls across multiple high-risk jurisdictions — particularly combined with AIS gaps at key junctures — are consistent with evasion routing. Repeated operation in zones associated with sanctions-driven transshipment (certain anchorages off the UAE, Socotra corridor, Strait of Malacca approaches) should trigger enhanced underwriting scrutiny, regardless of declared cargo type.

Geographic risk profiling based on historical port call data allows underwriters to score vessels against peer group behaviours — identifying outliers whose routing patterns deviate materially from legitimate trading comparables.

‍

‍

War risk pricing: From corridor premiums to vessel-level behavioural scoring

The Red Sea crisis demonstrated both the limits and the potential of data-driven pricing. When Houthi attacks escalated in late 2023, war risk premiums for Red Sea transits surged from around 0.3% to as high as 0.7% of insured value. Initially, pricing was corridor-based: if you transited the Red Sea, you paid the premium.

As data accumulated, the market became more sophisticated. Insurers began identifying targeting patterns — Houthi preference for vessels with Israeli, UK, or US nexus — enabling more differentiated vessel selection and pricing strategies. The market effectively bifurcated: vessels with those touch points were priced differently from those without.

This is behavioural underwriting in practice. The next stage is extending that methodology beyond active conflict zones to the broader risk landscape — using vessel-level behavioural data to generate dynamic, continuously updated risk scores rather than applying blanket corridor premiums.

‍

Operationalising maritime intelligence in the underwriting workflow

The challenge for most underwriting teams is not awareness of these data sources — it is integration. Maritime intelligence is only actionable when it is embedded into the underwriting workflow at the right decision points.

There are three practical integration points:

1. At new business submission: Pre-bind screening should include a behavioural flag review covering AIS transmission history over the prior 12 months, ownership structure and any recent changes, port call sequencing against high-risk geographic indicators, and any documented STS activity or AIS manipulation events.
2. At mid-term: Continuous monitoring between renewals allows underwriters to detect material changes in vessel behaviour — new trade routes, ownership changes, AIS anomalies — and take appropriate action before loss events occur. This is where the static model fails most visibly: risks change between snapshots, and insurers are often the last to know.
3. At renewal: Behavioural data should form part of the renewal risk assessment alongside claims history, enabling underwriters to distinguish between vessels whose risk profiles have deteriorated (and price accordingly) and those with consistently clean behavioural records (and reward them competitively).

‍

The compliance-underwriting convergence

One of the most important structural trends for marine underwriters is the convergence of compliance risk and insurable risk. Historically, sanctions compliance was a legal and back-office function. Today, an insurer who unknowingly provides cover to a vessel engaged in sanctioned trade faces direct enforcement exposure.

In July 2024, OFAC sanctioned Ascent General Insurance Company for providing insurance to vessels tied to a Houthi financial network. This was not a minor sanction. It was a signal that the provision of insurance services to sanctioned maritime networks is itself a sanctionable act — and that due diligence obligations extend to underwriters, not just shipowners and operators.

Cargo theft compounds the picture: 2024 saw a 27% increase in cargo theft incidents compared to 2023, with average losses exceeding $202,000 per incident. Organised crime is adapting tactics, including exploiting fragmented documentation in complex multi-modal supply chains. Behavioural data — particularly container tracking intelligence and port dwell time anomalies — is increasingly relevant to cargo underwriting decisions.

‍

Where this leads

The market direction is clear. As one senior US marine insurance executive put it: "The industry is trying to use data to get away from reactive remedies and be proactive to avoid the claim in the first place."

For marine underwriters, behavioural-based risk modelling is not a technology project. It is a risk discipline — one that requires integrating continuous operational signals into every stage of the underwriting lifecycle. The tools to do this exist. The data is available. The regulatory environment is demanding it.

The underwriters who will sustain competitive advantage in this market are not those who can most efficiently process yesterday's loss data. They are those who can read the behavioural signals that precede tomorrow's losses.

‍