Motor Truck Cargo · Data Intelligence Platform
The data to price cargo risk correctly has always existed. It was just never in one place.
Motor truck cargo is one of the most complex property & casualty lines to price. Losses happen through crashes, yes — but also through heavy braking, spoilage, seal failures, and fictitious pickups. Fleetidy compiles everything that matters — public safety data, application intelligence, and claims outcomes — into a single platform where actuarial analysis can build a model that sharpens over time.
Layer One · Public Data
800,000 carriers. Seven federal datasets. A decade of safety history.
The FMCSA publishes complete operating history on every interstate motor carrier in the country. Census records, inspection reports, roadside violations, crash incidents, BASIC scores, authority history, and operating status. Fleetidy compiles all of it, normalizes it against fleet size and mileage exposure, and applies empirical weights validated against actual loss correlations. This is the baseline every carrier gets scored against before a single application is filed.
Crash-per-Power Unit
Normalized crash frequency adjusted for fleet size and mileage exposure
Behavioral Risk Index
Weighted violation portfolio across 30+ violation types with empirical relative-risk multipliers
FRED Peer Index
Percentile rank within fleet-size peer group — a single number from 0 to 100
Beyond Crashes
Cargo is lost in many ways. You need to see all of them before you know which ones are predictable.
Crashes are the most visible failure mode — and FMCSA data captures them well. But cargo is also lost through heavy braking, spoilage, broken seals, fictitious pickups, fire, and dozens of other causes. The only way to know which loss types have predictive triggers is to collect claims data across all of them. Once you have the full picture, the data itself reveals which causes can be managed through underwriting decisions — terms, exclusions, pricing, declination — and which are essentially random perils. For the loss causes that turn out not to be predictable, you still need to know their loss cost per unit of insurance to confirm that the rate is adequate.
Loss causes Fleetidy captures
Crashes
Collisions, rollovers, jackknifes. Already well-covered by Layer 1 FMCSA data — crash frequency, violation history, and safety scores.
Inertial
Hard braking, rapid acceleration, sharp turns — cargo shifts or is damaged without any collision. Requires loss-type data from applications.
Adulteration
Spoilage, rusting, contamination, broken seals, refrigeration failure. Highly commodity-specific.
Theft & Fraud
Cargo theft, fictitious pickups, hijacking. A rising loss category correlated with lane, commodity value, and operational profile.
Fire & Weather
Vehicle fires, severe weather events, natural disasters. May or may not have carrier-specific predictors — the data will tell.
Other Perils
Mechanical failure, warehouse incidents, and any other cause. Captured so the full loss cost per unit of insurance is known.
The principle: collect claims data across every loss type. The data reveals which causes have predictive triggers — those get built into underwriting rules. The causes that turn out to be unpredictable still matter: their loss cost per unit of insurance sets the floor for rate adequacy. Both outcomes require Layers 2 and 3.
The Architecture
Three data layers. Each one builds on the last.
Predictive pricing requires more than a snapshot. It requires a model that accumulates institutional knowledge over time — what risks were submitted, what was written, what was declined, what claims resulted, and how those outcomes compare to the initial risk assessment. Fleetidy is structured as three compounding data layers.
Layer 1 · Public Data
ActiveFMCSA census, inspections, crashes, violations, BASIC scores, and authority history for 800K+ interstate carriers.
Normalized, scored, and available at point-of-quote. No application required.
Layer 2 · Application Data
Grows With Every SubmissionStructured data from every application — written and not-written. Both the risks you bound and the risks you declined contain pricing signal.
Commodities, lanes, coverage limits, loss run data, premium history, and reasons for decline.
Layer 3 · Claims Outcomes
Closes The LoopOne year after binding, what actually happened? Claims frequency, severity, and loss type fed back into the model validate and sharpen Layer 1 and Layer 2 signals.
This is the feedback loop that makes predictive pricing possible.
Institutional Knowledge Accumulates
Every submission adds a data point. Over time, the platform holds a complete picture of what you have seen, what you have written, and why. That picture is searchable, analyzable, and exportable.
Actuarial Analysis Without Manual Extraction
Instead of pulling spreadsheets from three different systems, the data is already normalized and joined. Actuaries and data scientists work directly on clean, structured output — with or without ML tooling.
Each Cycle Sharpens The Next
As claims outcomes arrive, the model's predictive correlations are validated or corrected. Variables that show actuarial lift get elevated. Variables that show no signal get deprioritized. The model improves with every underwriting cycle.
Submission Management · Layer 2 Collection
Every application is a data asset. Treat it like one.
The submission pipeline is the primary interface for collecting Layer 2 data. Every application — whether quoted, bound, referred, or declined — contributes commodity profiles, coverage structures, loss run histories, and operational context that feeds the pricing model. Dual-client workflows ensure both broker intent and underwriter reasoning are captured, not just the final outcome.
Broker Dashboard
Create applications with structured data capture — commodities, lanes, coverage needs, and loss history. Submit to the underwriting pool when ready. Track every submission through quoting, binding, and issuance.
Underwriter Dashboard
Claim submissions from the incoming pool. Access the carrier's FRED Score alongside application data at point-of-quote. Rate, quote, refer, or decline — every decision is recorded with structured reasoning.
Shared Messaging
Threaded conversations between broker and underwriter, directly on the submission record. No context lost. No forwarded chains. Every message tied to the account it belongs to.
File Uploads & Auto-Classification
Upload or email documents into any submission. Files are automatically sorted by type — loss runs, certificates, binders, photos — so the submission file stays organized without manual tagging.
Tiered roles ensure brokers submit, underwriters quote, managers oversee their teams, and admins control the platform. Referrals are structured with reason codes and tracked. Every action is auditable.
Security · Protecting Institutional Data
The value is in the data. The security model reflects that.
Fleetidy accumulates proprietary pricing intelligence — loss histories, decline reasons, commodity correlations, and carrier assessments that do not exist in any public dataset. The security architecture is designed to protect that institutional knowledge with enforced access controls, full audit trails, and insurance-grade session management.
Two-Factor Authentication
TOTP-based 2FA with QR provisioning and backup codes. Enforced at login and available for all user roles.
Role-Based Access Control
API-level enforcement of role permissions. Brokers cannot access underwriter actions. Managers see only their team. Admins have a complete audit view.
Full Audit Trail
Every login, submission, status change, quote, and document upload is logged with timestamps, user IDs, and IP addresses. Searchable and filterable by category and date range.
Secure Session Management
Bcrypt password hashing, CSRF protection, secure cookie configuration, and session expiration policies. XSS-safe output escaping throughout all templates.
Underwriting Workspace · Where Data Becomes Decisions
Everything an underwriter needs. One record.
When an underwriter claims a submission, they get a tabbed workspace with the full picture: application details, the carrier's FRED Score and safety profile, messaging with the broker, private notes, referral tools, a rating calculator, and a quote builder. Every decision — quote, refer, decline — is recorded with structured reasoning that feeds Layer 2.
Integrated Rating
Enter trucking revenue, storage revenue, and rate percentage. Premium calculates automatically. Push directly into a formatted quote.
Structured Referrals
Refer by capacity, authority concern, risk complexity, or specialty. Set priority, designate the recipient, and add context notes. The referral trail stays on the record.
Carrier Intelligence Map
Every submission links to the carrier's full FMCSA profile — geospatial HQ mapping, inspection timeseries, safety metrics, and the FRED Score — right on the Insured Info tab.
Decline with Transparency
When a risk does not fit, decline with a documented reason that is sent directly to the broker. No ambiguity. No lost follow-ups. The decline reason becomes a data point in Layer 2.
Quoting · Policy Issuance
Quotes that read like policies. Generated from rated forms, not text boxes.
When an underwriter issues a quote, Fleetidy generates a policy-grade PDF — declarations, coverage forms, schedule of endorsements, signature lines — assembled from a curated motor truck cargo form library that automatically resolves the right state variant for the insured’s mailing state. On bind, a Binder PDF is produced from the same form set so brokers can deliver compliant evidence-of-coverage same day.
State-Aware Form Resolution
Mailing-state schedule auto-resolves the right form variant — no manual lookup, no compliance gaps.
Endorsement Library
Theft, refrigeration, trailer interchange, electronic equipment, parked-unattended, contingent cargo — all wired to forms.
Owner’s Cargo Mode
Private-carrier and hybrid scenarios get a dedicated owner’s-cargo form set — first-party direct damage handled cleanly.
Auto-Generated Binder PDF
When underwriting binds, the Binder PDF is produced automatically and attached to the submission for broker delivery.
Negotiation · Drift Detection
Counter-offers, captured. Stale quotes, flagged.
Brokers can counter underwriter quotes with revised premium, limits, deductible, or endorsement stack. Underwriters can re-quote, accept, decline, or reverse a counter. Each round is preserved with side-by-side red-box diff highlighting. When the underlying application changes after a quote is issued — exposure update, new vehicle, revised commodity profile — the platform marks the quote stale so the underwriter knows to re-quote before bind.
Multi-Round Negotiation Threads
Counter, re-counter, accept, reject, withdraw, reverse — every round timestamped and tied to the user who took the action.
Side-by-Side Diff Highlighting
Premium changes, endorsement adds/drops, limit moves — every delta is visually called out so reviewers can see what shifted.
Stale-Quote Detection
If the application is updated after a quote is issued, the quote is automatically flagged stale until the underwriter reviews and re-quotes.
Quote Expiration & TTLs
Quotes expire on a configurable TTL (default 30 days). Expired quotes block bind until refreshed; brokers see the countdown.
Clearance & Binding · UA Workflow
Two pools. One pane of glass.
Underwriting Assistants run a two-pool dashboard. The Clearance Pool catches duplicate submissions across producers, surfaces conflicts, and records grant/block/unblock decisions with rationale. The Binding Pool handles post-bind handoff — binder issuance, policy numbering, document delivery — so nothing falls between underwriter and broker.
Cross-Producer Conflict Detection
Two producers shopping the same insured? UA sees both and routes ownership cleanly — no double-quoting, no broken broker relationships.
Grant / Block / Unblock
Every clearance decision carries a reason and an auditable trail. Reversible — if circumstances change, unblock with a note.
Binding Pool Handoff
When underwriting binds, work flows to UA for binder issuance, policy numbering, and claim setup. Status stays visible to the broker.
Two-Pool KPIs
Pending clearance, open conflicts, cleared count, binder backlog — the dashboard surfaces queue health at a glance.
Layer 3 In Action · Claims & Loss Runs
FNOL to settlement, with loss type captured at every step.
When a claim arrives, FNOL captures structured loss-type data — crash, inertial, theft, adulteration, fire — that feeds Layer 3. Investigation notes build a timeline; reserves split indemnity, expense, and ALAE; settlements close the file. Every claim is tagged back to the bound submission, so frequency and severity flow into the actuarial cohort rate book. Loss runs from carriers and brokers are uploaded as PDFs and parsed into structured frequency/severity records attached to the submission.
Structured FNOL by Loss Type
Every new claim is filed against a loss-type taxonomy. The data starts clean — no after-the-fact reclassification needed.
Investigation Timeline
Notes, photos, adjuster activity, supplemental documents — all on a single timeline keyed to the claim.
Reserves & Settlement
Indemnity, expense, and ALAE tracked separately. Payments and close events recorded with audit history.
Loss Run PDF Parsing
Drop in a loss run PDF, get structured rows out. Attach to a submission so frequency/severity flow into Layer 2.
Claims-to-Cohort Feedback
Every closed claim joins back to the bound policy’s cohort, validating or correcting the rate book.
Claims Dashboard & Team Roster
Open vs closed pipeline, adjuster assignments, manager oversight — built for a real claims org chart.
Actuarial · New-Business Pricing
A cohort rate book that reads your bound book and writes your next quote.
The actuarial dashboard groups every bound policy into cohorts by fleet-size band, FRED Grade, and cargo tier — then publishes suggested low / mid / high rates for each cohort using your own loss experience. When an underwriter rates a new submission, the platform pulls the matching cohort’s published range as a guideline. A tiered fallback (most-specific cohort → broader cohort) ensures sparse cohorts still have data backing them. Shock losses can be excluded with reason codes; loss types can be reclassified after FNOL. Every rate movement is auditable.
Cohort Portfolio Grid
Fleet size × FRED Grade × cargo tier — a heatmap showing exposure, premium, claims, and loss ratio for every cell of your book.
Published Rate Book
Suggested low/mid/high rate per cohort, surfaced into the underwriting workspace at the moment of quote.
Tiered Fallback
Specific cohort sparse? Fall back to the broader band, then broader still — with the fallback level shown to the underwriter.
Shock Loss Exclusion
Outlier losses can be excluded from cohort math with a documented reason. Auditable, reversible, visible in the rate report.
Loss Reclassification
After FNOL, if a loss is recategorized (e.g. theft → fictitious pickup), the cohort rate is recomputed with audit trail.
Closed Loop with Layer 3
Every settled claim feeds the matching cohort’s experience. The rate book sharpens with every cycle.
Management & Senior Oversight
Funnel metrics, hit rates, and loss ratios — by underwriter, by grade, by cohort.
Managers see what their team is shopping, quoting, and binding — and how each underwriter’s grade distribution compares to the team’s. Hit rate by FRED Grade reveals where the book is winning; premium per power unit shows rate adequacy; loss ratios surface book-quality drift in time to correct it. Senior management gets cross-circle visibility — a single pane across multiple teams without compromising tenancy.
Pipeline Funnel
Submission → quoted → bound → issued, per underwriter and per team. Drop-off at each stage is visible.
Hit Rate by FRED Grade
A-grade conversion vs C-grade conversion — which segments of the book are winning, which are bleeding.
Premium Adequacy
Premium per power unit by cohort against the published rate book — instant visibility into rate slippage.
Loss Ratio Analytics
Trailing 12-month and inception-to-date loss ratio by cohort, grade, and underwriter.
Underwriter Comparison
Side-by-side underwriter benchmarking on grade mix, hit rate, average premium, and loss ratio.
Senior Cross-Circle View
Manager II sees across multiple circles for portfolio-level oversight — without breaking team-level tenancy.
Supporting Capabilities
A platform built to be operated.
Communication Hub
Threaded in-app messaging tied to every submission. Inbound email routed via webhook into the right thread. Push notifications (APNs + web push) with distinct chimes per type. In-app help desk and guided onboarding for every role.
Mobile (iOS Native + PWA)
Native iOS app with Face ID / Touch ID login and APNs push for underwriters in the field. PWA fallback runs anywhere with service-worker-backed offline browsing of the carrier database.
Admin & Operations
Circles isolate organizations multi-tenant. Trash & recovery prevents data loss from accidental deletes. Encrypted backups protect institutional data. Smoke tests, server logs, and hermetic dev/prod environments keep operations transparent.
Public Research & Validation
The methodology is published. The data is validated.
Predictive pricing only works if the underlying model is testable. Fleetidy publishes the complete scoring methodology — component weights, violation classifications, EB shrinkage, eligibility gates — and backs it with empirical validation studies that compare scores against actual loss outcomes. Cross-validation, ROC, and AUC studies establish discrimination. Commodity-specific (hazmat) and domicile studies surface subgroup signals.
Methodology
Component weights (56/18/14/12), EB shrinkage parameters, grade cutoffs, eligibility gates — fully documented.
Empirical Validation
Score vs actual loss outcomes — does each grade band predict the loss frequency it should?
K-Fold Cross-Validation
Robustness testing across folds — confidence the model isn’t overfit to a particular slice.
Pool AUC / ROC Study
Discrimination metrics — how well does the score separate carriers that crash from carriers that don’t?
Hazmat Risk Study
Commodity-specific subpopulation analysis — is the score equally predictive on hazmat risks?
Domicile Study
Geographic subpopulation analysis — does the model hold up across state operating-base profiles?
Start building your pricing model
Your first data layer is ready. Start compiling the rest.
The FMCSA foundation is already scored and waiting. Sign up to begin adding your application data, your underwriting decisions, and your claims outcomes to a system designed to turn institutional knowledge into predictive pricing.