Enterprise AI
Model governance
Processes for approving, documenting, monitoring, and retiring models.
Example
Model governance tracks validation, approval, monitoring, and retirement for deployed models.
Why it matters
It helps prevent forgotten or unsuitable models from continuing to influence decisions.
You'll hear it
“Is this version still approved for the task we're using it for?”
Can you control it?
Sometimes
Sometimes. An organization can define or influence model governance through its policies, processes, contracts, and product choices. Individual users may have fewer controls.
Common questions
- Does approval last forever?
- Not necessarily. Changed data, uses or performance can justify another review.
- Does governance end when training finishes?
- No. Monitoring, updates and retirement remain part of the ongoing work.
- Can a model be retired while its records are retained?
- Yes. Historical information may still be needed to understand earlier decisions.