You're being asked to cover AI risk — D&O, PI, cyber, the EU AI Act liability regime — off a questionnaire the applicant filled in themselves, that nobody can verify and that's stale the day it's signed. The Insurable AI Standard gives you a live, Bitcoin-sealed proof of controls you can verify in seconds, re-check any time, and that demonstrably lowers your loss ratio.
Every decision is sealed and provable after the fact.
Logged, and inputs are provenance-checked - garbage-in is caught.
Logged, sourced, tools shielded and a human is accountable.
Fully underwritable: nothing autonomous moves money without a sealed sign-off.
Pillars marked VERIFIED are checked live against a sealed proof; ATTESTED pillars are declarations for you to spot-check. You still set the premium — sebbi gives you the evidence, not the price.
Paste an applicant's code (IC-XXXX-XXXX). You get the level, the pillar evidence and the Bitcoin seal — no account needed.
Add one line to your AI risk appetite: "AI deployments must hold a current sebbi Insurability Certificate at Level I-3 or above." Every insured then has to certify — and the proof is one you can check yourself, not take on trust. Adopt it self-serve in two minutes: declare your minimum level, prove it by domain control, and get a certificate-gate your pipeline calls on every submission. Verifying publishes a sealed announcement.
Become a founding insurer →Who already requires it →Talk to us first →Machine-readable spec & API →The standard is open. A certificate records provable controls, not a guarantee against loss; the premium is yours to set. Browse the public register →