Inside an ocean cargo submission: what InsightXtract extracts — and why it matters

Ocean cargo is deceptively hard to intake. The submission is usually just two documents — the broker’s email with the ask, and a marine cargo application — but the risk lives in the details: what is being shipped, how it moves, where it goes, and how much is exposed on any one vessel at any one time. A crate of machine parts moving port-to-port on a container line is a different risk from temperature-sensitive pharmaceuticals routed through a war-risk corridor, even at the same annual values. The information that decides the price is scattered across the application’s cargo profile, its conveyance schedule, and its loss history — and it is almost never on any one page.

Most intake tools read one document at a time and hand back a flat list of fields. That is not how a marine account is underwritten. InsightXtract runs an agentic, multi-document pipeline that classifies each file, extracts to a per-document schema, consolidates everything into one unified account record by source priority, and validates it against your rules and reference data — so the underwriter opens a single, coded, cited record instead of an email and a PDF.

One submission, two documents, one connected record — classify, extract, consolidate, link, validate.

The idea that governs everything: the exposure is per conveyance, not per year

Before the categories, the concept that makes marine cargo different. In most lines you rate an annual aggregate; in ocean cargo, the exposure that hurts you is what sits on a single vessel or shipment. Annual shipping values tell you the size of the book of goods moving, but the limit that matters is the limit per conveyance — the maximum the carrier is on the hook for if one ship sinks, one container is lost overboard, or one warehouse burns. That is why the application carries both an annual values figure and per-conveyance and per-location limits, and why the conveyance schedule breaks the exposure down route by route.

So InsightXtract captures both dimensions and keeps them distinct. In the tables below, the Captured column flags how each parameter is read:

• Current — captured once for this term.
• Point-in-time — a single static fact about the account.
• Per-row — a table: one route/conveyance, or one claim, per row.
• Value base — an exposure figure that anchors the rate.

A · Insured & identity

Who the account is — legal identity, classification, and the operating profile that frames the cargo risk. Sources: marine cargo application, broker email.

Parameter	Captured	Why it matters
Insured name · DBA · FEIN · entity type	Point-in-time	The contract party and unique account key — drives clearance, conflicts, and de-dupe.
Address · risk state	Point-in-time	Domicile and filing jurisdiction; the risk state is standardized against a reference table.
NAICS code · SIC code · business description	Point-in-time	Class-based appetite and hazard grade — an importer, a manufacturer, and a freight forwarder ship very differently.
Years in business	Point-in-time	Stability and operating-experience signal for the account.

B · Size of the operation

The financial scale behind the cargo — the numbers that size the account and sanity-check the shipping values. Source: marine cargo application.

Parameter	Captured	Why it matters
Estimated revenue / annual sales	Value base	Overall account size — a reasonableness anchor for the values being shipped.
Total payroll	Value base	Operational footprint and workforce scale of the insured.
Employee count	Point-in-time	Size and complexity of the organization handling the goods.

C · Cargo profile

The heart of an ocean cargo risk — what moves, how much of it, and by what modes. Susceptibility to damage, theft, and spoilage all start here. Source: marine cargo application.

Parameter	Captured	Why it matters
Commodity · commodities shipped	Current	The single biggest susceptibility signal — electronics attract theft, perishables spoil, steel corrodes, chemicals escalate severity.
Annual insured values	Value base	Total goods at risk over the term — the primary exposure base for premium.
Annual shipping values	Value base	The throughput of goods moving — frequency of exposure across the year.
Conveyance · conveyance types	Current	Mode of transit (vessel, air, truck, rail, intermodal) — each mode carries its own loss profile.

D · Coverage & limits

The exact shape of the cover being requested — and where marine cargo diverges sharply from other lines. Per-conveyance and per-location limits, war/SRCC, and the valuation basis are the parameters that move the price. Sources: marine cargo application, broker email.

Parameter	Captured	Why it matters
Limit per conveyance	Current	Maximum exposure on any one vessel/shipment — the core capacity decision for a cargo layer.
Limit per location	Current	Accumulation exposure while goods sit in a warehouse, port, or storage — a distinct catastrophe risk.
Deductible	Current	Retained loss per event — the frequency layer the insured absorbs.
War / SRCC coverage	Current	War, strikes, riots & civil commotion — a separately underwritten peril tied to geopolitical routes and corridors.
Valuation basis	Current	How a loss is valued (e.g. CIF + 10%, invoice, replacement) — it directly sets the amount payable.
Requested limit	Current	The capacity the broker is asking for — the headline of the ask.
Premium	Current	The price on the table — the rate baseline for quoting and comparison.

E · The conveyance schedule — one row per route/mode

Where the exposure actually lives. InsightXtract reads the conveyance schedule as a table — one row per route and mode — so the underwriter sees the risk lane by lane rather than as one blended number. Source: marine cargo application.

Column	Captured	Why it matters
Conveyance type	Per-row	The mode for this lane — ocean vessel, air, truck, or intermodal, each with its own loss frequency.
Route	Per-row	The trade lane — origin/destination determines transit time, piracy/war corridors, and port-handling risk.
Limit	Per-row	The exposure carried on this route — per-lane accumulation the underwriter must price and aggregate.
Commodity	Per-row	What moves on this lane — ties susceptibility to route so hazard is evaluated in context.

F · Loss history — one row per claim

The single biggest pricing input. InsightXtract reads each claim as a row from the application’s loss history, with the paid/reserve/incurred split that shows both cost and open development. Source: marine cargo application.

Column	Captured	Why it matters
Claim number	Per-row	The loss identifier — de-dupes and links a claim across documents.
Date of loss	Per-row	When it happened — drives the frequency trend and loss-year analysis.
Description	Per-row	Nature and cause of the loss — water damage, theft, general average, shortage, contamination.
Status	Per-row	Open or closed — whether more development is still possible on the claim.
Paid	Per-row	Cash already out the door on the claim.
Reserve	Per-row	Money still set aside — the open exposure that can develop upward.
Incurred	Per-row	Total cost (paid + reserve) — the number that feeds loss ratio and severity.

G · Broker & submission

Who’s placing the account and the terms of the ask — the distribution chain and the transaction envelope. Source: broker email.

Parameter	Captured	Why it matters
Broker name · broker email	Point-in-time	Distribution routing, binding authority, and correspondence.
Product	Current	The line/coverage form being requested — workflow and appetite routing.
Effective date	Current	Term start and binding deadline — standardized to a canonical date format.

From two documents to one connected record

Pulling these parameters out of an email and an application is only half the job. The value is in consolidation: the insured named on the application and the insured on the email, the requested limit in the email and the per-conveyance limit on the app, the commodities in the cargo profile and the commodities on each schedule row — they all describe one account. InsightXtract merges them by a declared source-of-truth priority — the application wins on cargo, coverage, and loss details; the email carries the broker and the ask — and every value is cited back to the document and page it came from.

On top of the merge, the pipeline runs reference-data validation: the risk state and NAICS code are standardized and checked against lookup tables, currency fields are normalized, dates are formatted, and required fields (like the insured name) are enforced — so what reaches the underwriter is not just extracted, it is coded and clean.

Why it matters to the business

Comprehensive, structured, route-aware extraction isn’t a data-entry nicety — it changes the economics and quality of the book:

• Route and conveyance exposure, made visible. Capturing the conveyance schedule row by row — mode, route, limit, commodity — means lane-level accumulation and war/piracy-corridor risk reach the underwriter instead of hiding inside a blended annual figure.
• Commodity susceptibility, priced properly. Electronics, perishables, chemicals, and steel fail in different ways. Extracting the commodity on the account and on each schedule row lets the model treat susceptibility as the pricing driver it is.
• Per-conveyance limits, not just annual values. The loss that hurts is what sits on one vessel or in one warehouse. Keeping limit-per-conveyance and limit-per-location distinct from annual values is the difference between rating throughput and rating catastrophe.
• War / SRCC, flagged by default. War, strikes, riots & civil commotion is a separately underwritten peril tied to geopolitics. Extracting the war/SRCC term automatically means it is a deliberate underwriting decision, not an afterthought.
• Consistency and auditability. The same categories, coded the same way, every time — risk state and NAICS validated against reference data, every value cited to its source. That is a repeatable book, not one that depends on which underwriter opened the file.

The Ocean Cargo agent extracts all of this today — the insured and identity; the size of the operation; the full cargo profile of commodities, annual and shipping values, and conveyance modes; the coverage terms including per-conveyance and per-location limits, deductible, war/SRCC, and valuation basis; the broker and submission; and both the conveyance schedule and the loss history as tables — consolidated into one coded record, every value cited to its source document and validated against reference data. And because it’s all configuration — fields and tables in the agent’s output contract, not code — the schema keeps pace with what marine underwriters ask for.