Ask your AI vendor for the evidence the NAIC Bulletin expects

Why ask

The NAIC Model Bulletin, adopted by the National Association of Insurance Commissioners in December 2023, sets out what state regulators expect from insurers using AI. Each state decides whether to adopt it. As of 31 August 2026, the NAIC's adoption map shows 26 jurisdictions (25 states and the District of Columbia) that have adopted it, with California, Colorado, New York and Texas listed separately for their own insurance-specific rules or guidance.

The Bulletin is clear that buying AI in does not move the responsibility. An insurer's AI programme should cover how it acquires and relies on AI systems developed by third parties, including due diligence on the vendor, contract terms such as audit rights and co-operation with regulators, and audits or other checks on the vendor's compliance.

This request turns those expectations into twelve plain questions for the vendor. The answers are self-declared, so treat the result as a list of documents to ask for, not as proof.

How it works

1. You create a request link above and send it to your vendor from your own email.
2. Your vendor answers 12 questions on this page about the evidence they hold.
3. You both get the same result: what is available, what is partly available and what is missing, with the rule behind each item and a fingerprint so anyone can check it was not changed.

The 12 questions

1. Intended use and objectives: A written description of what the model is meant to do, its goals and objectives, and how it was developed and validated to meet them. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, predictive models narrative
2. Development and use documentation: Detailed documentation of how the model was developed and how it should be used. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, model documentation
3. Third-party data used to build it: A list of the external data sources used to develop the system. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Third-Party AI Systems and Data
4. Data currency, lineage and quality: Records of data currency, lineage, quality, integrity and suitability for the use. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, data practices
5. Bias analysis and minimisation: Results of bias analysis on the data and model, and what was done to reduce it. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, data practices
6. Validation and testing: Validation, testing and retesting results, including how outputs generalise once implemented. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, validation and testing
7. Drift, robustness and reproducibility: Assessments of model drift, robustness, repeatability, reproducibility and traceability, and how they are kept up to date. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, model assessments
8. Human involvement in decisions: A description of how far people are involved in the final decision when the system is used. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Risk Management and Internal Controls, degree of human involvement
9. Support for consumer notice: Information the insurer can use to tell affected consumers that an AI system is in use. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): General Guidelines, notice to impacted consumers
10. Audit rights or audit reports: Contract terms giving the insurer audit rights, or audit reports from a qualified auditing entity. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Third-Party AI Systems and Data, contract terms
11. Co-operation with regulators: A commitment to co-operate with the insurer on regulatory inquiries and investigations about its use of your product. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Third-Party AI Systems and Data, contract terms
12. Evidence the insurer can use for due diligence: Material that lets the insurer confirm decisions supported by your system meet the legal standards that apply to the insurer itself. NAIC Model Bulletin, Use of AI Systems by Insurers (2023): Third-Party AI Systems and Data, due diligence

Sources

• NAIC Model Bulletin: Use of Artificial Intelligence Systems by Insurers (adopted 4 December 2023)
• NAIC: implementation of the AI Model Bulletin (adoption map)

Sources checked 8 October 2026. This tool helps you ask for evidence. It is not legal advice, an audit or a certification, and the answers are the vendor's own.