how it works · the front door
audit-gated.quoted in minutes.
we are a carrier, not an auditor. if you hold a recognized ai agent audit, you are in scope and we quote fast. if you do not, we point you to auditors and pick up once you are certified. keeping those two roles separate is what makes our paper neutral.
tell us about your agent
a short intake: what it does, how much autonomy it holds, its financial authority, and its blast radius.
hold a recognized audit?. yes
in scope. upload it, get terms.
your audit maps to our intake, its tier sets your control credit, and we return an indication in hours. no re-audit, no duplicate work.
hold a recognized audit?. no
referred out, then welcome back.
we hand you to recognized third-party auditors. we never run the audit ourselves. come back certified and we bind.
priced on authority & blast radius
premium keyed to what the agent is allowed to do unattended, not to which model powers it. drop the authority, drop the price.
repriced on what actually happened
optional telemetry feed from your agent logs. renewal tracks real behavior, not a stale questionnaire, and earns a trace discount.
we do not price the model
we price what the system is permitted to do
autonomy tier
from human-approves-everything to full autonomy with sub-agents or unsupervised operation.
authority ceiling
the largest transaction an agent can execute unattended; the force, speed, and workspace envelope a machine can operate in near people.
irreversibility
can the action be undone? a clawable refund and a recoverable config are small risks; a deleted database and a broken arm are not.
reach
how many customers, records, systems, or square meters one bad decision touches.
rate
how many actions can repeat before a circuit breaker trips.
premium is a transparent build-up on those factors, with control credits for the safeguards that demonstrably shrink them: caps enforced in code, circuit breakers, geofencing and speed limits, rollback, kill-switches. plus a governance credit up to 20% for passing certification. two companies running identical hardware can pay very different premiums, and the difference is entirely in their own hands.
segments & indicative structure
| segment | typical limits | indicative premium |
|---|---|---|
| sme deploying off-the-shelf agents | $1m – $5m | $5k – $40k |
| ai vendors & big startups (agents) | $10m – $50m | $75k – $500k |
| enterprise multi-agent deployers | $25m+ | $250k – $2m+ |
| drone & amr operators | $1m – $10m | usage-based, from low thousands |
| humanoid & robot fleet programs | $5m – $25m | bespoke, capacity-partnered |
planning-grade indications pending filed rates and capacity terms. not an offer of insurance.
hard gates, not soft discounts
the exclusion list is enforced as binary underwriting gates, not premium loadings. a system that fails a gate is declined or sub-limited with written conditions precedent to fix, never quietly priced up. this is what keeps a young carrier solvent, and it is what a capacity partner most wants to see.
fast, parametric-leaning claims
wherever a loss is measurable from data, we pay from data: refund-anomaly spikes, resolution-rate collapse, fleet downtime windows. parametric triggers settle in days. if the telemetry proves the event, the payment should not wait for a quarter of correspondence.
audits we recognize
what goodfault does,and deliberately does not
one job done well: carry the risk. everything here supports that. the audit, pointedly, is not on the list.
agentic liability coverage
the core product. brokered on specialist paper today, goodfault-issued as our binding authority lands.
instant quote for the certified
hold a recognized audit and get terms in minutes, not weeks. the certificate is your fast lane.
open audit recognition
we accept any credible ai agent audit against published criteria. we are not tied to one auditor. and never our own.
continuous underwriting
telemetry-fed pricing that reprices on real agent behavior at renewal, with credit for the trace feed.
the ai damage report
our free public registry of real agent failures with loss estimates. the market's incident record, and our pricing edge.
auditor referral network
not certified yet? a warm handoff to recognized auditors, so the path to coverage is one introduction away.
everyone else audits the risks they underwrite. your auditor should not also be your insurer, for the same reason your accountant should not audit their own books. goodfault is the neutral carrier that makes the whole market's certificates worth something.
like what you see?
get covered.
tell us what your agent does, what tools it can reach, and what a bad day looks like. we return a premium in plain english. no actuarial latin, no 40-page pdf. most policies bind in under a week.
get a quoteconsult the oracle
not ready for a quote? describe your agent to the oracle and get an honest read of your riskiest failure mode. the one your roadmap is politely ignoring. free, anonymous, occasionally prophetic.
