02 — Capability · BC-1780.50
Model Risk Management
Know which models drive risk, pricing, capital and financial reporting decisions, prove each one fit for purpose before and during use, and limit the harm a wrong model could do, including AI models in regulated use. AI policy sits in AI Strategy & Governance, and building and running machine learning models in AI & Automation.
- Model Governance
- Model Risk
- MRM
In scope
- Model inventory, ownership and materiality tiering, including AI models in regulated use
- Independent model validation and findings
- Ongoing model performance monitoring
- Model change approval and use restrictions
Out of scope
- Building, deploying and operating machine learning models (see BC-770.20)
- Responsible AI policy across the enterprise (see BC-770.10)
Realized by · 0
- No product in the catalog yet.
Used in · 1
Build it · 0
- Nothing in the library points here yet.
Decomposes into · 4
- BC-1780.50.10Model Inventory & TieringKeep a complete inventory of models with owners, uses and dependencies, and tier each by materiality and complexity.
- BC-1780.50.20Independent Model ValidationTest a model's design, data, implementation and outcomes independently of its developers and record findings that must be fixed.
- BC-1780.50.30Model Performance TrackingTrack each model's accuracy, stability and overrides in use and trigger revalidation when it degrades.
- BC-1780.50.40Model Approval & Use LimitsApprove new models and material changes before use, and set limits or overlays when a model is used with known weaknesses.