Agent Risk Score Stability Assurance — Pilot Program Service Contract

IMPORTANT NOTICE: THIS AGREEMENT IS A SERVICE CONTRACT AND LIMITED WARRANTY. IT IS NOT A CONTRACT OF INSURANCE.

1. Nature of Agreement

This Agent Risk Score Stability Assurance Plan ("The Plan") is a service contract designed to manage the unique risks associated with the operation of Synthetic Actors and autonomous AI Agents. This Plan is ancillary to, and distinct from, any Third-Party Liability or Professional Errors & Omissions (Tech E&O) insurance policy you may hold.

2. Scope of Protection: The "Stochastic Gap"

Standard software insurance protects against negligence and coding errors (deterministic failures). This Plan specifically covers Stochastic Failures—instances where an AI Agent, operating on correct code and infrastructure, makes a probabilistic decision that results in an error, hallucination, or unintended action.

We cover:

  • Semantic Hallucinations: Generation of false, misleading, or defamatory content by the Agent.

  • Unintended Transactions: Autonomous execution of buy/sell/delete commands outside of the Agent’s intended parameters.

  • IP Infringement: Unintentional generation of copyrighted material by the Agent.

3. Benefit Limit: Immediate Defense Retainer

Upon verification of a Trigger Event (defined below), Agent Risk Score will provide a fixed, immediate payment to the Subscriber.

  • Benefit Amount: $2,000.00 (USD) per verified incident.

  • Payout Speed: Funds are wired within 72 hours of claim verification.

  • Purpose: This benefit is designated as an "Emergency Legal Retainer" to allow the Subscriber to immediately engage legal counsel for a response to a Cease & Desist or Demand Letter.

4. Trigger Events

A claim under this Plan is triggered when the Subscriber receives a written Formal Demand (e.g., Cease & Desist letter, Lawsuit Summons) from a third party that specifically alleges financial loss or damage caused by the actions or outputs of the Subscriber's AI Agent.

5. Subscriber Obligations & Cognitive Telemetry

To maintain eligibility for this Plan, the Subscriber agrees to:

  1. Primary Insurance: Maintain an active Technology Errors & Omissions (Tech E&O) insurance policy with an A-rated carrier throughout the term of this Plan.

  2. Cognitive Telemetry (Deferred): The Subscriber agrees to install the Agent Risk Score Risk Monitoring SDK upon its public release. This SDK provides the forensic data (Cognitive Telemetry) required to distinguish between code bugs and stochastic hallucinations. Failure to install the SDK within 30 days of release may void this warranty.

  3. Stability Index: The Subscriber agrees to remediate Agents that fall below the critical threshold of the Agent Risk Score Stability Index.

6. Exclusions

This Plan does not cover:

  • Deterministic Errors: Standard software bugs, server outages, or downtime (covered by your standard Tech E&O).

  • Intentional Malice: Actions taken by the Agent at the specific, prompted instruction of the Subscriber to cause harm.

  • Criminal Acts: Any illegal activity conducted by the Subscriber.

7. Legal Disclaimer

This Plan is provided by Agent Risk Score Inc. It is not guaranteed by any state insurance guaranty association. This Plan serves as a warranty on the performance stability of the AI Agent verified by Agent Risk Score’s infrastructure.