destill.ai · Insurance · Trust Center

For underwriters of cyber, E&O and AI liability

Seven signed numbers per insured company. Do they say anything about claims?

Status 3 November 2026: a plan, no data yet. Whether these metrics predict claims or losses is unproven. This pilot is designed so that it can show they do not. No price, premium or score is derived from them before it has run.

Many cyber, E&O and liability policies cover damage caused by AI without pricing it ("silent AI"). From 9 December 2026 the new EU Product Liability Directive treats software, AI included, as a product. Our firewall gateway already counts, per insured company, what its AI did and signs the count. This page says how an insurer pilot tests whether those counts carry risk information.

The seven metrics

One signed report per insured company and month, from the AI gateway in front of its models. Hashes and counts only: no prompt, no answer text, no personal data.

idmetricwhat it countswhat it does not show
U1AI steps with a verified proofagent actions in signed POAW records that verify, divided by all agent actions in those recordsagents that were not wired to POAW
U2AI calls through the firewallevery call of the month that passed the gatewayAI use that bypasses the gateway
U3Attacks blocked at the inputcalls stopped before the model saw them, also per 1,000 callsa count, not a detection rate: false alarms are in it, missed attacks are not
U4Leak eventsanswers that disclosed confidential instructions, stopped in the stream or delivered and judged afterwardsleaks the checks do not recognise
U5Locked keysrequests refused because a key was paused after a leak, plus keys revoked in the monthwhether a revocation was an incident
U6Agent self-reports of fabrication or driftsigned agent runs that flagged fabrication or drift from their taskagents that do not report
U7Failed or aborted callscalls that did not complete as askeda loss: this is availability

An underwriter checks each report without trusting us: the signature against our published key (offline possible), a recount of the numbers from the rows the signed report itself lists, and a fresh request for the same month that must give the same numbers. Changing one figure makes the check fail and name it.

How the pilot is set up

Hypotheses, and what would refute them

#hypothesissupported ifrefuted if
H1Months with a leak event (U4 > 0) are followed by more data-disclosure incidentsrelative risk ≥ 2, 95 % interval above 1 (assumed threshold, fixed in advance)95 % interval upper bound below 1.5 (assumed threshold)
H2Input-attack pressure (U3 per 1,000 calls) rises with manipulation incidentsrank correlation ≥ 0.3, interval excludes 0interval upper bound below 0.1
H3Self-reported fabrication or drift (U6 > 0) is followed by more complaint incidentsrelative risk ≥ 2, 95 % interval above 1 (assumed threshold, fixed in advance)95 % interval upper bound below 1.5 (assumed threshold)
H4With high proof coverage (U1 ≥ 0.9) the cause of an incident can be settled from the records alone≥ 70 % of incidents vs ≤ 30 % without coveragedifference below 15 points
H5Control: failed calls (U7) do not predict incidentsinterval includes "no effect"U7 predicts as well as U3, U4 or U6

The idea as a whole is refuted if none of H1 to H3 holds, or if the control H5 predicts incidents as well as they do. Then the metrics only measure how much AI traffic there is, not risk. H4 concerns claims handling and can hold on its own.

How many company-months

Assumptions, to be replaced by the insurer's own base rates before the start: incident rate 5 % of company-months without the signal and 10 % with it, signal present in one month out of five, 5 % significance, 80 % power. That needs about 270 company-months with the signal and 1,090 without, roughly 1,360 in total, for example 115 companies over 12 months. With fewer, the result is published as a description only, with no statement about prediction.

Notified claims are much rarer than incidents. A pilot of this size cannot test claims directly; they are reported, but conclusions rest on incidents. Claims need pooling across insurers or a longer run.

Who gives what

Back to insurancePilot onboarding