What it means to automate transport underwriting with an AI layer
To automate transport insurance underwriting with AI is to place an external AI layer on top of the insurer's existing systems that reads each cargo submission, enriches it, scores route and cargo risk, proposes a premium, and either decides clean risks or escalates the rest to a human underwriter, all calibrated to the insurer's own risk appetite and underwriting manual. The core stays in place. The intelligence sits above it. This matters most in the Transport line (ramo transportes), where the same cargo on two different routes is effectively two different risks, and where a submission arrives as a mix of broker email, PDF proposals, fleet spreadsheets, risk-management plans, and prior loss runs.
The Transport line is one of the oldest and most technical lines inside Brazilian Seguros e Danos (P&C, property and casualty), and it grows because the economy moves on trucks. Underwriting here is administratively heavy by nature. Roughly 40% of an underwriter's time goes to administrative tasks rather than risk decisioning, according to Deloitte, which is precisely the load an AI layer is built to reclaim. The reader who should care is the underwriting (subscrição) lead, the product and innovation head, the C-level deciding whether automation is feasible without an IT project, and the broker (corretor) who wins cargo business on response speed.
How the automated journey works in transport insurance
The automated journey follows the same logic a Transport underwriter already uses, but the AI layer carries the repetitive load and hands the human a decision-ready file. It runs in six stages. First comes multichannel intake with automatic validation, where the submission enters in the format the insurer already receives, through email, attachments, upload, or API, with no change to how brokers submit. Second is intelligent document reading, where Machine Learning extracts the structured fields a Transport underwriter needs from unstructured documents: cargo type and declared values, the routes and highways involved, the risk-management plan, carrier data, and loss history.
Third is broker enrichment and context, where the submission is cross-referenced against external and internal sources such as carrier history already on file, route exposure, accumulation against the existing book, and broker conversion history. Fourth is the risk and fraud engine, a multi-factor ML model that produces a risk score for the specific cargo, route, and operation. High-theft cargo on a high-theft corridor with weak risk management scores very differently from the same cargo on a controlled route with escort (escolta) and tracking. Fifth is dynamic pricing, where the layer calculates a risk-adjusted premium (prêmio) and conditions consistent with the insurer's pricing logic and the Transport underwriting manual, including loadings tied to risk-management requirements. Sixth is decision and prioritization: clean, in-appetite risks can be quoted or straight-through processed, out-of-appetite risks are declined or routed to a human underwriter with the full enriched file and the rationale, and the result writes back to the policy core with a complete audit trail.
How to deploy the external AI layer in transport
Deployment is structured so the insurer never runs an IT project and never migrates its core. WIR is 100% external, an AI layer on top of the systems the insurer already operates, with no load on the insurer's IT team. The work begins by scoping the Transport journey as it exists today, mapping how submissions arrive, which coverages are written, and where underwriters lose time. From there the layer integrates with the existing core and policy systems through the channels the insurer already uses, so the system of record stays exactly where it is.
The decisive step is calibration. The model is tuned to the insurer's own underwriting manual for Transport, not to a generic template. Risk-appetite rules for cargo types, maximum value per shipment, accepted routes, and required risk-management controls are encoded so the score and any automated decision reflect the insurer's stated appetite. Manual-driven loadings, escort thresholds, and value limits map to the same conditions the underwriting team applies today. This setup phase is a one-time engagement that runs 3 to 12 months, with fixed scope and KPIs agreed before start, followed by continuous operation in production after go-live. Because Transport risk moves with highway theft dynamics, the appetite and scoring are recalibrated over time rather than left static.
Governance, explainability, and LGPD
Automating an underwriting decision in Brazil sits inside two regimes, and governance is a design requirement, not an optional feature. Under the LGPD (Lei Geral de Proteção de Dados, Lei 13.709/2018), Article 20 gives the data subject the right to request review of decisions taken solely on automated processing, which means an automated underwriting decision must be explainable and reviewable. The national authority is the ANPD. Separately, SUSEP supervises the P&C market, including pricing adequacy and conduct, so automated pricing must remain technically justifiable.
The practical requirement that follows is auditability. Every automated score, price, and decision must be traceable to its inputs and to the insurer's encoded appetite, so the insurer can answer both a customer review request under the LGPD and a regulator question under SUSEP. The AI layer is built for exactly this. Every decision is explainable and returns a full audit trail, data is encrypted at every step, and the layer logs its rationale per decision. The insurer keeps control of the tail because only clearly in-appetite risks are decided automatically, and everything borderline or outside appetite is escalated with context to a human underwriter. This is the difference between automating data work and automating decisions the insurer cannot explain.
How WIR automates transport underwriting
WIR is the AI layer for insurance, an external intelligence platform that automates the quotation and underwriting journey on top of the insurer's existing systems, calibrated to the insurer's risk appetite and underwriting manual. In the Transport line this means the WIR layer reads the cargo submission, enriches it, scores route and cargo risk, proposes pricing, and decides or escalates, while the insurer's core remains the system of record. WIR is not an insurer, a broker, or an MGA, and it does not carry risk. It automates the journey; the insurer's appetite governs what gets bound versus escalated. The AI layer for insurance. On top of the systems the insurer already runs, never in their place.
Two modules carry this work. Underwriter Intelligence automates the quotation journey per the insurer's risk policy, with real-time ML scoring calibrated to appetite, automatic routing by appetite and exposure, and predictive conversion analysis by product, risk, and broker, so underwriters spend their time on the risks that need judgment. Smart Sales adds distribution intelligence, mapping the portfolio and scoring next-best-action, which matters because 60%+ of brokers choose an insurer by response speed, according to Capgemini. Real-time dashboards give a proactive view of in-flight deals. WIR's relevant public traction for this line is a first POC in execution with a global insurer in the Transport line. It is a proof of concept in progress, stated conservatively, not a signed client and not a named customer. Every automated decision remains explainable, auditable, LGPD compliant, and encrypted at every step.
Frequently asked questions
Does automation account for route, cargo, and transport exposure?
Yes. The risk and fraud engine scores the specific cargo, route, and operation together, because the same cargo on two routes is effectively two different risks. A multi-factor Machine Learning model weighs cargo type, declared value, the highways involved, the risk-management plan, escort and tracking, and accumulation against the existing book. High-theft cargo on a high-theft corridor with weak controls scores very differently from the same cargo on a controlled route.
Does the AI layer replace the core in the transport line?
No. WIR is 100% external, an AI layer on top of the systems the insurer already runs, never in their place. The core stays the system of record. The layer integrates through the channels the insurer already uses, with no IT project and no migration. It reads the submission, enriches, scores, prices, and decides or escalates, then writes the result back to the policy core with a full audit trail.
How is the model calibrated to the transport underwriting manual?
The model is tuned to the insurer's own Transport underwriting manual, not a generic template. Risk-appetite rules for cargo types, maximum value per shipment, accepted routes, and required risk-management controls are encoded, so the score and any automated decision reflect the insurer's stated appetite. Manual-driven loadings, escort thresholds, and value limits map to the same conditions the team applies today. Because highway theft dynamics shift, the appetite and scoring are recalibrated over time.
Are transport decisions explainable and auditable?
Yes. Every automated score, price, and decision is explainable and returns a full audit trail traceable to its inputs and the insurer's encoded appetite. This lets the insurer answer a customer review request under the LGPD, Article 20, and a SUSEP question on pricing adequacy. Data is encrypted at every step. Only clearly in-appetite risks are decided automatically. Everything borderline or outside appetite is escalated to a human underwriter with full context.
Does WIR have experience in the transport line?
WIR's relevant public traction for this line is a first POC in execution with a global insurer in the Transport line. It is a proof of concept in progress, stated conservatively, not a signed client and not a named customer. WIR is the AI layer for insurance, and the Transport line is where it automates the cargo quotation and underwriting journey, calibrated to the insurer's risk appetite and underwriting manual.