Model Governance
Policies, controls, and oversight processes managing the full lifecycle of predictive and AI models from development through retirement.
technicalPublished 2026/06/07Last verified 2026/06/07
FAQs
- Does model governance apply to vendor-supplied models we did not build ourselves?
- Yes. Regulatory guidance consistently holds that an insurer is responsible for models it uses, regardless of whether a third party developed them. You should require the vendor to provide documentation equivalent to what you would produce internally and perform your own validation before deployment.
- How often should we revalidate a deployed pricing model?
- Most governance frameworks require at minimum an annual revalidation and a triggered review whenever material changes occur in the book composition, a data feed changes, or monitoring metrics breach defined thresholds. High-frequency models used in real-time scoring may warrant quarterly monitoring reviews.
- What is a model inventory and is it required by regulators?
- A model inventory is a centralized register listing every model in production along with its owner, purpose, validation status, and next review date. While no single federal mandate requires it in insurance, NAIC guidance and state market conduct examiners increasingly expect carriers to produce one on request.
Related Terms
Model Risk Management
A framework for identifying, measuring, and mitigating risks from quantitative models—including pricing models, fraud scores, and AI systems.
AI Model Audit
A structured review of an AI or statistical model's design, training data, outputs, and deployment to verify accuracy, fairness, and regulatory compliance.
Algorithmic Bias
Systematic unfair discrimination in AI or ML models disadvantaging protected classes—a critical compliance concern as insurers adopt predictive models.
Model Drift
Degradation of a deployed model's predictive accuracy over time as input feature distributions or outcome relationships shift from the training environment.