Designed to be insurable: the AIA Guard insurability program
Insurance for AI agents starts with proof. AIA Guard’s built-in insurability program turns every assessment into a policy-ready dossier: the complete evidence pack a future insurer needs to understand the risk. Cover itself always comes from an insurer’s own policy — our job is to make your agent ready to be insured.
Your policy-ready dossier
The dossier separates verified observations from applicant statements and lists missing evidence openly. It stays versioned and time-bound, so a future reader can tell whether the insured system, workflow and permissions still match the reviewed setup.
| Dossier area | Evidence recorded | Question it answers for an insurer |
|---|---|---|
| Use and scope | Agent version, workflow, users, languages, tools and exclusions. | What activity and consequence are being considered? |
| Data exposure | Data types, processing locations, providers, access paths and retention. | Which confidential or personal data could be affected? |
| Authority | Tool scopes, approval points, customer-account boundaries and stop controls. | What can the agent do without another person? |
| Verification | Sample design, failures, retests, pillar results and blockers. | Which controls were actually tested, and with what limits? |
| Operations | Incident history, monitoring, change log and responsible owners. | How will an important change or failure be detected and handled? |
| Remaining risk | Open findings, excluded cases, dependencies and review date. | Which uncertainty remains after the assessment? |
How the insurability track works
The track runs inside every assessment: we define the insured activity precisely, test the controls that contain it, and document the rest — openly. The result is a dossier written for reuse: procurement and governance teams use it today, and an insurer can build its own review on it tomorrow. That is what “designed to be insurable” means.
We are equally open about the road ahead: no insurer partnership has been signed yet, no policy, premium, limit or eligibility rule exists, and the draft score thresholds are not yet tested against real claims data. The dossier makes no promise an insurer must keep — it gives the insurer clear reasons to say yes.
Who decides what
You own the accuracy of submitted facts, the operation of the agent and notification of incidents or important changes. AIA Guard tests one bounded setup and states its limits. The insurer alone would decide underwriting, wording, limits, exclusions, premium and cover under a real policy.
A prospective buyer should ask how model, data, tool or autonomy changes affect the dossier; whether known incidents or excluded workflows must be disclosed; which losses, services or third-party dependencies a policy would exclude; what limits, deductibles and notification duties apply; and whether a changed agent needs renewed underwriting. No answer is implied by an AIA Guard score — the dossier exists so those questions get precise answers.
A clear regulatory boundary
FINMA explains that offering or concluding insurance — including advice aimed at concluding a contract — can count as insurance intermediation, while providing data or information alone is distinguished. This site stays informational: it prepares evidence, it never sells or promises cover. Any future distribution model would need its own legal and regulatory review.
Common questions
Does an AIA Guard assessment include insurance cover?
No — by design. The assessment produces a policy-ready dossier about one tested setup. Cover exists only under an insurer’s own contract, wording, limits and exclusions.
Which insurers accept the dossier?
We are building the dossier to the standard insurers need, before insurer conversations begin. No insurer or broker agreement has been signed, and none is claimed.
Will a strong result lower our premium?
We promise no discount or pricing effect. A strong, well-documented result gives an insurer clear reasons to say yes — the decision, and the price, remain entirely theirs.