Skip to main content

For insurers and underwriters

You are already writing AI risk. You just cannot see it.

Somewhere in your book are insureds whose consequential decisions are made by systems nobody in the business can explain, with no named human accountable for the outcome. Nothing on a proposal form asks.

And you know the shape of what comes back when it goes wrong. The system made the call, nobody understood it well enough to be responsible for it, so no individual can be held to it. AIC exists to make that answerable — and verifiable by you, without taking anyone's word for it.

To be plain about where this stands: no insurer currently recognises AIC certification, and no organisation has been certified yet. This page sets out what the signal would be and how recognition would work. We would rather propose it than describe it as though it already exists.

The signal

What you have today, and what this adds

AI governance reaches you the way most non-financial risk does: as the insured's own account of itself. The gap is not that organisations lie on proposal forms. It is that nothing on the form can be checked.

Source
TodayThe insured, describing themselves
With AICAn independent assessment against published requirements
Basis
TodayA tick box, or a paragraph of prose
With AICEvidence an assessor examined, graded against 44 published requirements
Scope
TodayRarely defined — “we use AI responsibly”
With AICNamed systems and decision types, stated on the record
Verification
TodayNone available to you
With AICDirect from AIC, no login, without asking the insured
Currency
TodayTrue on the day it was signed, if then
With AICLive status — suspensions and lapses show on the record
Accountability
TodayUsually a department
With AICA named individual who has signed a personal declaration

What you would get back

One field on a proposal form, answerable in thirty seconds

No login, no account, no call to us. A certificate number resolves to a record that states its own scope and its own expiry, and shows plainly when something has lapsed or been suspended.

A badge that cannot visibly lapse is a badge that means nothing. The status history stays on the record — a revoked certificate is marked revoked, never quietly deleted.

aiccertified.cloud/verifyIllustrative record
Certified, activeAIC-D3-2027-0041
Division
D3 Reviewed — AI decides, humans review a defined sample
Scope
Retail credit origination and collections decisioning
Accountable Person
Named on the certificate record
Issued
14 February 2027
Expires
14 August 2028, on the 18-month cycle for D3
Continuous monitoring
Live — telemetry coherent with declared Division
Status history
No suspensions, no revocations

Illustrative. AIC has issued no certificates — the public register is empty and says so. This is the shape of the record, not a real one.

How verification works

Confirm a status directly from AIC — never from the insured

1

An insured presents an AIC badge

Every AIC certificate will carry a QR code and a verify URL — printed on the badge, not just claimed in a proposal.

2

You confirm it directly with AIC

The verify page answers in under 30 seconds, with no login: status, scope, Division and expiry.

3

The status is never taken on trust

A suspended or lapsed certification shows plainly, with its history — a badge that can visibly lapse is a badge that means something.

Verify a certificate

What the mark verifies

An evidence-based audit, not a self-declaration

AIC certification is an evidence-based audit against published requirements, mapped to the regulatory frameworks that apply to the certified organisation. It confirms that a named individual is accountable for the organisation's AI-driven decisions, that an override process exists, and that the certification is checkable — not just claimed.

A certification that can quietly lapse without anyone noticing isn't worth much. AIC is developing Pulse, continuous monitoring between audits. When it is available, certified organisations whose Pulse telemetry is live and coherent will be able to carry a Continuously Monitored overlay — a mark that stays accountable after the audit, not only on the day of it.

AIC also runs a free, self-declared tool called AIC Aware. It carries no independent verification, never appears on the certified register, and is not something an underwriter should price against — that is what the audited mark is for.

Between certifications

The record itself, if the insured lets you read it

A certificate is a point-in-time judgement. The record behind it is kept every day, in the AIC platform: the insured's AI systems and who answers for each, the decisions they make and how often a person overrides them, and evidence AIC reads nightly from the systems the insured has connected.

An insured can issue you a key from their workspace. With it you read an extract of that record directly from AIC, whether or not they are certified yet.

What it contains
Observations only: counts, coverage, rates and dates, each traceable to something on the record.
What it leaves out
Any rating, grade or recommendation. Pricing and acceptance are yours, not AIC’s.
Who controls it
The insured. They issue the key, and they can withdraw it.
Why it can be trusted
It is built from the same source as the insured’s own overview, so you never see a different picture from the one they manage.

Recognition

Naming who recognises the mark — not what AIC does for you

Where an insurer has agreed to treat AIC certification as a factor in underwriting, the certificate's verify page names that directly:

AIC certified, recognised by [insurer name] as an underwriting consideration

Deliberately not “for premium benefit”. Whether verified AI accountability correlates with insurable risk is an open question, and the evidence to answer it does not exist yet. That is the question worth testing together — not a claim to make in advance of it.

What AIC does not do

AIC certifies governance, not products, and does not conduct assessments on behalf of an insurer or for insurance purposes. AIC does not price risk, underwrite, or advise on coverage. An AIC certification is one input an insurer may choose to use — the decision, and the terms, remain the insurer's alone.

The direction matters: insurers recognise AIC certification — AIC does not certify for insurance. That line is what keeps the certification body impartial, and it is not open to blurring.

Testing the signal

The useful question is whether this predicts anything

AIC is not asking any insurer to endorse a certification. It is building an independent, evidence-based measure of AI accountability, and the open question — whether it carries real underwriting signal — can only be answered with insurers rather than at them.

If that is a question your team finds interesting, we would like to talk about how it could be tested: what would need measuring, what AIC would provide, and what would count as an answer either way.

Start that conversation