RouteRisk: Risk as a Property of Future Route State
The mobility and insurance industries have spent twenty years attributing risk to the driver. Telematics scores, trip reports, hard-braking events, fatigue indicators — all of it is a proxy for the question: how risky is this person? RouteRisk starts from a different premise. Risk is not what a driver does. It is what a route will be when the vehicle gets there.
That shift in attribution — from person to route state — is not semantic. It changes what data matters, what signals are predictive, and which organisations are the natural buyers. A driver score tells you about history. A route state score tells you about the next six hours of exposure, before a wheel turns.
Risk is not what a driver does. It is what a route will be when the vehicle gets there.
What the Current Stack Cannot Answer
Three categories of tool currently serve this space. Each handles one dimension well. None of them answer the operational question that matters most: how risky is this specific route, right now, given all interacting conditions?
Telematics platforms
e.g. Samsara, Geotab
Monitors: Driver and vehicle behaviour
Gap: Route risk is invisible until after an incident occurs
Weather intelligence
e.g. WeatherOptics, Tomorrow.io
Monitors: Environmental conditions by location
Gap: No model of how weather interacts with vehicle type, load, or route timing
Insurance pricing
Carriers, MGAs, Lloyd's syndicates
Monitors: Historical exposure and loss runs
Gap: No real-time situational input; pricing reflects the past, not current route state
No existing tool combines all three dimensions into a forward-looking route-level risk score.
Dispatch decisions are made without real-time risk input. Insurance pricing lacks spatial and contextual precision. When a claim occurs, there is no objective situational record of what the route state was at time of incident. RouteRisk is built to close all three gaps simultaneously.
Risk-Weighted Miles
RouteRisk introduces a new operational unit: risk-weighted miles. Not miles driven. Not miles driven safely. Miles driven under a quantified, time-indexed, vehicle-specific risk score — computed before the vehicle departs, recomputed continuously as conditions evolve.
The underlying model treats the road network as a dynamic risk field. Every segment carries a hazard rate — a function of road structure, weather forecast, traffic state, vehicle type, and historical incident density. A route is a sequence of segments. Route risk is the sum of segment hazard rates weighted by exposure. The system finds the route that minimises that sum within operational constraints: delivery window, distance, driver hours.
Risk-Weighted Miles — The Core Unit
Segment hazard rate
λ(s, t, v) = f(Rs, W(s,t), T(t), Vv)
Expected accident exposure per unit distance for a given vehicle type, road segment, and time window — adjusted for current environmental conditions.
Route-level risk score
Risk(route) = Σ λ(sᵢ, tᵢ, v) · E(sᵢ, tᵢ)
Segment hazard rates summed across all route segments, weighted by traffic exposure — producing a comparable, dispatch-ready risk score for any route alternative.
Who This Is Built For
RouteRisk sits at the intersection of three professional domains. Each has a distinct need — and a distinct reason why the signal is immediately legible.
Underwriter
Adjusts premium or declines route class
Dispatcher
Selects route with lowest expected loss
Driver
Receives selective alerts — not spam
Commercial Fleet Insurance Underwriters
Per-mile risk pricingSpecialty MGAs — Canal Insurance, Protective Insurance, Lloyd's trucking syndicates
These underwriters already price per-mile exposure. They will immediately recognise a route-level risk score as an additive signal — one that captures the environmental and structural conditions their actuarial models cannot see. Real-time route state data enables more precise per-mile pricing, active loss prevention, and a structured claims record built automatically at time of incident.
Loss Control Consultants
Claim reduction toolsLarge brokers — Marsh, Aon, Lockton
Loss control teams sit between insurers and fleets and are measured on claims reduction. They need tools that are operationally deployable — not research outputs. RouteRisk plugs into existing dispatch workflows and delivers actionable routing guidance. A prevented loss is a direct win for the broker relationship and the carrier's combined ratio.
Autonomous and Semi-Autonomous Fleet Operators
The 2025–2035 hybrid windowWaymo Via, Aurora, Kodiak — and the platforms following them
In the hybrid autonomy era, accident liability is migrating from drivers to fleets, OEMs, and software systems. Route state intelligence becomes essential infrastructure — both for safe operation and for demonstrating due diligence to insurers and regulators. These operators need a risk layer that functions without driver behaviour as an input variable.
The 2025–2035 Hybrid Window
The timing argument for RouteRisk is structural, not cyclical. Three independent shifts are converging in the same decade.
↑ Severity
Severity rising
Accident frequency is declining as ADAS adoption grows. Severity is increasing — larger vehicles, higher repair costs, greater liability per incident. Insurers are pricing into a harder loss environment even as claim counts fall.
← Driver → System
Liability migrating
As automation increases, liability moves from the driver to the fleet operator, the OEM, and the software system. Route state becomes evidence — who knew what, when, about the conditions ahead.
→ Per-mile
Pricing evolving
Insurance is moving from annual aggregate pricing to usage-based, per-mile models. That shift requires a granular, real-time risk signal. RouteRisk provides exactly that unit of measure.
The 2025–2035 window is where complexity and ambiguity are highest. Mixed fleets — human-driven, semi-autonomous, and fully autonomous — operating on shared infrastructure, under insurance frameworks that were designed for a simpler liability model. RouteRisk is purpose-built for this environment.
Where RouteRisk Sits in the Stack
RouteRisk is not a data provider. It is a decision engine. The distinction matters for how it integrates and how it is sold.
Positioning in the Mobility Stack
| Category | Operates on | Output |
|---|---|---|
| Telematics (Samsara, Geotab) | Current vehicle/driver state | Visibility — what is happening now |
| Weather intelligence | Environmental conditions | Conditions — what the environment is doing |
| Insurance pricing | Historical exposure data | Premium — what past risk cost |
| RouteRisk | Future route state | Decision — what risk to accept before dispatch |
Commercial vehicle insurance is a $300 billion global market. Fleet SaaS is $10 billion. RouteRisk is positioned as intelligence infrastructure for the larger market — delivered as an API-first platform with per-vehicle, per-mile, and enterprise integration pricing.
We don't insure risk. We quantify and reduce it — before it happens.
The road already contains the next loss event. The route plan already exists. The only thing missing is the system that reads the intersection of the two — before the vehicle departs.
Concrete Example
Scenario — Reno to Oakland, Departure 18:00
Planned route
Refrigerated truck scheduled Reno → Oakland via Donner Pass, 18:00 departure.
Hazard detected
⚠ ElevatedWeather system over Donner Pass increases segment hazard rates by 40% within the planned arrival window.
Recompute
→ RecommendationRouteRisk evaluates alternatives. Southern deviation identified: +18 minutes transit time, −27% expected loss exposure.
Dispatcher accepts
✓ Decision madeSouthern route confirmed. Departure proceeds on schedule. No delay — only a route change.
Insurer notified
✓ LoggedReduced exposure logged pre-trip. Structured record created: original route, hazard state, recommendation issued, action taken.
+40%
Segment hazard rate increase
+18 min
Additional transit time
−27%
Expected loss exposure
Click any hashtag to explore related content on LinkedIn
Related Articles
RouteRisk.AI — The Future of Commercial Routing
How a graph neural network over a dynamic road hazard field converts structural, environmental, and operational data into real-time dispatch guidance.
Risk Map: The Future of Navigation and Risk Mitigation
Navigation tells you the fastest way to get there. Risk Map tells you whether getting there — on that route, at that time, in that vehicle — is a decision worth making.
Before the Vehicle Arrives: Predictive Risk Intelligence for Commercial Fleets - Waze for Fleets
How Trajectory Intelligence inverts the fleet safety model — evaluating risk along a planned route before a vehicle ever reaches a hazardous segment.
Goetz Weber
Contributor @ Maei.ai
Pioneer in motion intelligence, trajectory AI, and semantic routing. Theoretical physicist turned AI innovator.
Follow me on LinkedIn for insights on motion intelligence and AI
Share this article on LinkedIn with suggested hashtags:
Copies the article title, link and hashtags for you to paste into LinkedIn.