Trucking excess is auto-driven: the price turns on the fleet, the drivers, the radius, and a loss picture dominated by auto claims. The submission looks different from a construction account — a vehicle schedule instead of a WIP workbook, a driver roster, an auto-heavy loss run. We ran a real motor-carrier packet, Ironline Freight Systems, through InsightXtract; every result below is an actual output.

The packet & classification

Five files, routed to their document types:

Document
Classified as
Conf.
broker_email.pdfrenewal narrative + layer options
excess_casualty_submission_email
0.98
rfq.pdfmotor-carrier application
excess_casualty_application
0.97
fleet_exposure.xlsxvehicle schedule, drivers, WC, terminals
excess_casualty_exposure_workbook
0.99
loss_run.pdf5-yr Auto / GL / WC
excess_casualty_loss_run
0.98
schedule_of_underlying.pdfprimary + lead umbrella
excess_casualty_schedule_of_underlying
0.96

The fleet exposure workbook

The heart of a trucking submission. InsightXtract reads each sheet into a typed schedule — vehicle body types, license states and WC class codes normalized against glossaries. Sample of the vehicle schedule:

Unit #YearMake / ModelTypeGaragedCost New
U-21032023Freightliner CascadiaTractor vehicle_body_typesKansas City, MO us_state_codes$152,000
U-21182021Wabash DuraPlate 53'TrailerDallas, TX$48,000
… automobile 120 rows · driver_list 90 · workers_compensation 42 · location_list (terminals) 12

Header exposures fall out of deterministic derived rules over those rows — the numbers an underwriter rates on:

120
Power units + trailers
count(automobile)
90
CDL drivers
count(driver_list)
$56.7M
WC payroll
∑ workers_compensation.payroll
12
Terminals
count(location_list)

Vehicle-type split (tractor / trailer / straight truck) comes free from a group_by rule — the mix that drives auto rating.

The loss run — auto-driven

For a carrier, the loss run is the story. InsightXtract extracts every claim (coverage line, status and cause normalized) and rolls up incurred by line. This account’s 5-year, currently-valued picture:

31
Claims (5 yr)
count(claims)
$9.52M
Total incurred
∑ claims.incurred
$6.73M
Auto incurred
incurred_by_line · Automobile

Roughly 71% of incurred is auto — exactly what you’d expect and now quantified, tied to claim rows and a 10/01/2026 valuation date loss run p.1. (Note: a 31-claim table like this is where extraction JSON can truncate; the pipeline runs at a 64K output budget and retries with escalation, so the full table comes through.)

The tower

The schedule of underlying gives the attachment picture — primary Auto/GL/WC and the lead umbrella the excess sits over, coverage lines and admitted/E&S status normalized:

CoverageCarrierPer Occ.Status
Automobile Liability coverage_linesGreat West Casualty$1,000,000 CSLadmitted admitted_vs_surplus
Lead UmbrellaNorthland Insurance$10,000,000surplus lines
… 5 underlying policies; requested tower $10M / $25M / $50M

One coded motor-carrier record

Consolidated: fleet counts, driver count, WC payroll, auto-weighted losses, and the underlying tower — every value cited, every code normalized, every total deterministic. The same agent that reads a construction account reads this one; only the document types and glossaries do the industry work.