Claimy

Overview

Claimy is an AI-insurance startup that tests customer agents for failures and uses the results in insurance pricing.

Current Profile

Claimy’s model joins underwriting with technical risk reduction. It tries to provoke claim-relevant failures during onboarding, lets customers improve and retest their systems, and associates stronger scores with lower premiums.

Key Characteristics

• Uses adversarial agent testing to estimate failure risk before coverage.
• Tests concrete harms such as disclosure of sensitive customer information.
• Makes remediation and retesting part of the underwriting loop.
• Connects better test scores with lower insurance premiums.

Evidence

Failure discovery

• Insurers race to cover AI errors reports that one tested agent disclosed a previous customer’s age and personal profile in under ten minutes.

Price-linked remediation

• Insurers race to cover AI errors says failed agents can be improved and retested, with stronger results producing lower premiums.

Qualifications

• The episode does not disclose Claimy’s test suite, score calibration, false-positive rate, policy terms, customer count, or claims outcomes.
• A detected privacy failure demonstrates a vulnerability but does not by itself establish an actuarial loss probability.

What Changed

• Established Claimy as a test-linked AI insurance model.
• Added remediation and premium reduction as its central incentive loop.

Relationships

• Ines Butemacha - Claimy representative explaining its testing and pricing model.
• AI Agent Risk Testing - technical control at the center of Claimy’s approach.
• AI Insurance Data Scarcity - uncertainty Claimy’s tests attempt to reduce.
• Corgi - peer startup offering explicit AI liability coverage.