Home / Platform / Driving behaviour
The distribution case is interesting. The profitability case is the reason a reinsurer takes the call.
Continuous verified driving behaviour across a household portfolio, fed back to the driver every month, is a profitability instrument before it is a marketing feature.
Pilot
What is captured
Journey-level telematics through the handset: speed relative to limit, harsh acceleration and braking, cornering, phone handling while driving, time of day, and journey context. Two-wheelers included — a differentiator in markets where they carry a large share of motor exposure.
What is not captured: passenger identity, audio, or any location data outside the consented household circle. Capture is disclosed, consented in context, and can be withdrawn without losing the safety features.
The Driving Score
A normalised per-driver score built from exposure-weighted behaviour, designed so that a driver can see which journeys moved it and why.
Data ownership. The derived score is a contractual asset. Partner rights in it — use, retention, portability at termination — are set out explicitly in every agreement rather than left to a platform terms page.
The loss ratio chain
Six links stand between app engagement and a movement in the combined ratio. Each is stated with its current evidential status. An actuarial team should be able to interrogate this page and find nothing overstated.
- Evidenced
Capture
Handset telematics produce journey-level behavioural data at sufficient quality for scoring. Established technology, widely deployed.
- Evidenced
Score derivation
Behavioural data can be normalised into a stable per-driver score that discriminates between drivers. Established in the telematics market.
- Partial
Feedback loop
Monthly household reporting is opened and read at rates sufficient to influence behaviour. Evidence exists from adjacent products; Amanati’s own engagement rates are pending pilot.
- Partial
Behaviour change
Drivers receiving regular scored feedback improve measurably and sustain the improvement. Published evidence supports this in telematics programmes; the magnitude in the Moroccan market is unmeasured.
- Assumed
Claims effect
Improved behaviour reduces claims frequency and severity in this portfolio, at a magnitude material to the combined ratio. Directionally supported, not yet measured on Amanati data.
- Assumed
Portfolio self-selection
Drivers who opt into scoring are better risks on average, improving the book independently of behaviour change. Plausible and widely assumed; unquantified here.
Read this honestly. Two links are evidenced, two are partially evidenced, and two are assumed. Any partner building a business case on this chain should treat the final two as hypotheses the pilot is designed to test, not as findings.
Risk selection and pricing
Beyond behaviour change, the score is an input. It supports risk selection at new business, differentiated pricing where regulation permits, renewal retention of the best risks, and portfolio segmentation for the reinsurer.
Where local regulation restricts behavioural pricing, the score still functions for selection, engagement and retention. The regulatory position differs by market and is set out per deployment.
Monthly household report
Driving summary per driver, journey highlights, score movement and the specific behaviours behind it. Written for a family rather than an underwriter — which is why it gets opened.
This is the mechanism. A score nobody sees changes nothing.
Take the chain to your actuarial team.
The evidence pack sets out each link with its sources, its gaps, and the pilot design that would close them.