The actuarial record of AI-agent failure.

What Noxal is

Noxal is the agent loss registry: a structured, evidence-linked record of documented failures of AI agents in real deployments — classified to a consistent standard and scored for severity, so that the people who must price, govern, and underwrite AI risk can work from a common record rather than scattered headlines.

Every record in the registry is built from primary, publicly documented evidence — court rulings, regulatory filings, first-party disclosures, and reproducible security research — and is reviewed before it enters the feed. Where evidence is disputed or incomplete, the record says so.

Who it is for

Noxal serves insurance underwriters and actuaries pricing AI-related coverage, governance and risk officers assessing AI deployments, and the platforms that build AI risk into their own products. The classification standard is published openly; the classified record is provided by subscription.

How it works

Incidents are classified by failure class, root-cause layer, and autonomy level, and scored on the Agent Loss Severity (ALS) rubric — a transparent, reproducible measure. The methodology is open: any subscriber can reproduce any score from the record's own fields.

Contact

For subscriptions, partnerships, or press: info@noxals.com