For Insurers
Area risk with a confidence on every value, monitored continuously.
Per-dimension scores with confidence, not a black box. Save your weighting once, score every quote against it, and get a signed webhook the day a tracked area moves materially.
Built for underwriting
Four things a pricing team leans on.
Configurable scoring the actuary owns, monitoring that respects sample size, peer-relative screening, and delivery you can verify.
The composite the actuary owns
Save the weighting once as a preset. Every quote scores against it, per dimension, with a confidence on each.
Two sales never trip an alert
Moves are gated on transaction count. Thin evidence stays quiet instead of firing a false alarm.
Risky against its peer group
Rank areas by how far they sit from similar areas, not in absolute terms, so a real outlier stands out.
Signed, auditable delivery
Material moves fire an HMAC-signed webhook to your endpoint. Verify it before you process it.
Why it survives audit
Built to sign off, not to trust blindly.
Your model risk and actuarial teams ask how every number was made. The answer is in the response.
Versioned and published
Every score names the methodology version that produced it, and every release is published with the figures it moved.
Transparent components
Every score returns its dimensions, the weight applied and a confidence. Never a black box.
Sample-size honest
Moves are gated on transaction count. Static signals never fire, so two sales never trip an alert.
Frequently asked
Questions your pricing and model risk teams ask.
Can the actuary tune the weights without redeploying our codebase?
Yes. Save the weighting recipe once as a scoring preset and reference it by id on every quote. The actuary updates the preset from the dashboard and new quotes pick it up immediately. It is versioned, so model risk can track what changed and when.
How are the drift webhooks signed?
Stripe-style HMAC-SHA256 over the raw body, with the event type and a timestamped signature in the headers. The signing secret is shown once on subscription and never recoverable, so you store it and verify on receipt before processing. Endpoints must be public HTTPS.
What happens when the methodology changes?
Every response names the methodology version that ran, and every release is published in the changelog with the figures it moved. In a watched book, a change caused by a release is labelled a method change, not a move in the area, so your loss-ratio model is never surprised by one.
What do you store about our insured locations?
Just the postcode or area code and an optional label you supply to map back to your own policy id. No PII, no policy data, no premium amounts. The label is opaque to us.
Auditable, monitored, sample-size honest.
Score with weights the actuary owns, watch the book drift, and get a signed webhook the day a tracked area moves.