The risk score insurers and warranty providers price from.
Battery warranty is written today against a distribution nobody has seen. Reserves are set wide because the tail is unknown, and the cost of that ignorance is carried by every policy in the book. Cellworth prices the tail.
Three people who cannot do their job without this number.
Motor and warranty insurers
Writing battery cover on used EVs and reserving against a worst case they cannot bound.
Extended-warranty and RVI providers
Guaranteeing a residual value that turns almost entirely on a number they do not have.
Leasing and fleet finance
Setting end-of-contract values on assets whose principal component is unmeasured.
What actually lands in your hands.
- Risk score
- Probability that a pack falls below a chosen threshold within a chosen term — the shape of the tail, not a point estimate.
- Portfolio view
- The same distribution aggregated across a book, so reserves can be set against the portfolio rather than the worst vehicle in it.
- Monitoring
- Re-scored as new charge sessions arrive, so a deteriorating pack is visible before it becomes a claim.
- API
- A REST endpoint that returns a scored decision at quote time, in the time a quote engine has to spare.
Specimen · CW-0417-A
Price the policy instead of padding it.
An underwriter does not need certainty, they need a distribution they can defend to a capital provider. Give them a calibrated one and the loading that currently covers ignorance can come out of the premium — which is a competitive product, not a cost saving.
- 1 · Backtest
We score a closed cohort from your historical book and you compare our distribution against the claims you actually paid.
- 2 · Integrate
A quote-time endpoint, or a batch score against your existing book. No change to how you collect data.
- 3 · Reserve
Priced per scored policy, with the calibration report your actuaries and your regulator will both ask for.