We built a maritime intelligence platform. Here's why.

I spent the better part of last year watching marine underwriters work. Sitting in on calls. Shadowing how they evaluate a hull or cargo policy when the proposed voyage goes through, say, the Bab-el-Mandeb strait. The workflow was always some variation of the same thing: a Lloyd's List subscription in one tab, IMB piracy maps in another, the NGA warning list saved as a bookmark somewhere, a JWC circular from three weeks ago in their inbox, and a spreadsheet to tie it all together.

One underwriter — senior, 20+ years in London market marine — described it as "flying a 747 with a car dashboard." He knew more about Gulf of Aden risk than any dataset could tell him. But he was spending two hours per voyage assessment assembling data that should have been at his fingertips.

That's the gap. Not in the knowledge — underwriters are sharp. In the tooling.

---

The data fragmentation problem

Here's what a proper voyage risk assessment actually requires. I'm listing this not because you don't know, but because the sheer number of sources is the point:

Signal	Source	Update Freq.	Format
Piracy incidents	IMB ICC, ReCAAP ISC	Daily	RSS/XML feeds
War risk zones	JWC Hull Committee	Quarterly	PDF circulars
Nav warnings	NGA broadcast	Continuous	Plain text
Conflict intensity	GDELT Event Database	15 min	CSV dumps
Sanctions	OFAC SDN, UN, EU	Weekly	XML / CSV
PSC detentions	Equasis, Tokyo MOU	Monthly	HTML scrape
Flag performance	Paris/Tokyo MOU	Annually	PDF reports
Weather/sea state	Open-Meteo Marine	6-hourly	JSON API
AIS anomalies	AIS position feed	Real-time	NMEA stream
Country instability	Composite CII	Weekly	Computed

Ten sources. Ten different formats. Ten different update schedules. And the only "integration layer" most underwriting desks have is a person with good bookmarks and institutional memory.

The expensive platforms — you know the ones, six-figure annual contracts — give you some of this. They'll have piracy data, maybe weather. But they still make you alt-tab between four dashboards, and half the time the data is 48 hours stale. The cheaper tools give you one slice: piracy here, weather there, port state there. You're still the integration layer.

What underwriters actually need is one number they can defend in a peer review, with the receipts to show how it was computed.

So we built one

ArcNautical pulls all ten signals into a single composite voyage risk score. But the important word in that sentence isn't "composite" — it's "voyage." The score isn't for a port pair. It's for a specific route through actual water.

The distinction matters. A vessel going from Shanghai to Rotterdam can take Suez or go around the Cape. It can transit the Gulf of Aden hugging the Omani coast or track closer to Yemen. These aren't the same voyage. They have different threat exposures, different weather windows, different fuel costs. Scoring them as one number is wrong.

So we built a custom ocean pathfinder. Not a straight line on a Mercator projection — an actual A* search through a 0.05-degree resolution water grid, respecting coastlines, canals, and forced passages through narrow straits. The pathfinder produces a GeoJSON route, and then we sample every intelligence signal along that polyline.

Fleet monitoring, not fleet guessing

A voyage risk score at binding time is useful. Knowing what's happening to your fleet right now is essential. ArcNautical tracks vessels in real time via AIS, overlays every threat layer (JWC zones, active piracy, navigational warnings, conflict events), and alerts you when conditions change.

The alert system isn't a dumb geofence. You can configure it by signal type, severity threshold, vessel, or zone. When a vessel enters a JWC listed area, or a new piracy incident appears near an active voyage, or a vessel goes dark on AIS — the system notices and pushes the alert to your team. Webhook, email, dashboard notification.

Vessel vetting that follows the money

Every vessel has a registered owner. Most of the time, that owner is a single-purpose vehicle in the Marshall Islands. Who actually controls it?

ArcNautical pulls the GLEIF LEI registry, walks the corporate parent chain, cross-references against OpenOwnership beneficial ownership data, and screens every entity in the chain against OFAC, UN, and EU sanctions lists. You type in a vessel name and get back an ownership tree with sanctions hits highlighted and an opacity score that tells you how transparent — or opaque — the structure is.

Monte Carlo, because point estimates are cowardly

A voyage from Singapore to Hamburg takes roughly 25 days. That's what a deterministic model tells you. But weather changes. Sea states shift. A laden Suezmax handles Beaufort 6 differently than an empty Panamax. What if the monsoon shifts early? What if there's a 3-day weather hold at Suez?

We run 500 Monte Carlo simulations per voyage. Each simulation draws from probabilistic weather forecasts, applies vessel-specific speed curves, and correlates conditions across consecutive route segments (because a storm at waypoint 12 doesn't magically clear by waypoint 13). The output is a distribution:

"24.8 days" is the most likely outcome — but telling a charterer or underwriter that the voyage "takes 24.8 days" without mentioning the P90 is 27.6 days is presenting certainty where none exists. The distribution is the honest answer.

What's here now, and what's coming

ArcNautical is live. What works today:

• Voyage risk scoring — 10 signals, actual route geometry, active-signal weighting
• Fleet monitoring — real-time AIS, threat overlays, configurable alerts
• Vessel vetting — GLEIF ownership chain, sanctions screening, opacity scoring
• Stochastic modeling — 500-sim Monte Carlo with ETA/fuel distributions
• PDF reports — voyage risk, fleet summary, vessel dossier, piracy briefing, compliance
• Fuel and CII compliance — IMO DCS calculations, CII rating projections

Coming next: PML exposure analysis for portfolio-level risk aggregation. Deeper class society integration. Predictive threat modeling using GDELT event sequences. And better weather — real ensemble forecasts, not synthetic uncertainty.

---

If you're an underwriter, broker, ship manager, or fleet operator who's tired of stitching together risk data from eight different browser tabs — I'd like to hear from you. Not a sales pitch. Just a conversation about what's actually missing from your workflow. Every time someone in the industry tells me what's wrong with the platform, it gets better.