Research report · Version 1.1 · November 2026
What connected-vehicle data can and cannot tell an insurer, a broker or a fleet about road risk, and how that changes as automated driving moves from pilots into service. Covering the United States and Great Britain, with every claim graded by the strength of its evidence.
The six propositions
The report separates six claims that are often blended together, because they carry very different evidential weight.
Real-time intervention reduces commercial collision risk.Observed
Legitimacy now governs access to driving data.Observed
Outcome-trained, context-adjusted models outperform behaviour scorecards.Inferred
A severity-dominated fleet cannot evidence an intervention from its own claims.Observed
Automated driving will scale corridor by corridor, and the mixed fleet will define road risk well beyond 2030.Inferred
The largest open gap is a commercial-fleet layer that turns manufacturer data into claims-calibrated indices.Hypothesis
The verdict. The technology is ahead of the evidence, and the evidence is ahead of the market's contracts. Most of the value available before 2030 lies in the mixed human fleet, in proving which interventions move loss cost, and in structuring who captures the saving. In both worked fleets the intervention creates value overall but leaves the fleet that pays for it worse off over four years.
What is inside
- A three-layer model of vehicle data (aftermarket, manufacturer, video) and the governance tests each layer now has to pass in the US and GB.
- What actually predicts collisions, split into four separate questions, and a test protocol for any forecasting model before an underwriter relies on it.
- Two fleets with the same mileage and different loss profiles, worked through to a four-year cash flow for insurer, broker and fleet.
- Alert triage and video governance: how to run a programme drivers and regulators will accept.
- Liability and claims economics, including nuclear verdicts and the exoneration value of video.
- Automated driving: deployment evidence, what transfers between operating domains, and the economics of the mixed fleet.
- The market gap for a commercial vehicle-data layer, stress-tested.
- Actions for insurers, brokers and fleet operators that do not depend on any unproven claim, plus the evidence that would refute each proposition.
Contents
- Summary
- 1. Three layers of vehicle data
- 2. The governance question
- 3. What predicts collisions
- 4. Forecasting models and how to test them
- 5. Worked case: two fleets, the same mileage
- 6. Alert triage and video governance
- 7. Liability and claims economics
- 8. Automated driving
- 9. The market gap: a commercial vehicle-data layer
- 10. What can be done now
- Appendices A to F: vendor assessment, pilot design, business case, falsification conditions, source method, source audit
Choose an edition
Reader
£295 one person
- Full report, PDF
- Single named reader
- Instant download after payment
Full refund within 30 days if it is not useful to you. How refunds work
Team
£795 up to 10 people
- Full report, PDF
- Two-fleet business case workbook (Excel), with editable inputs and sensitivity
- Share within one organisation
Full refund within 30 days if it is not useful to you. How refunds work
Company
£1,500 whole organisation
- Everything in Team
- Company-wide licence
- A 60-minute briefing for your team on the findings
Prices exclude VAT where applicable. Evidence frozen 31 October 2026. The Connected Insurance Hub accepts no vendor sponsorship.