Govern which AI models are allowed

Settings, then Model governance, sets which AI models and providers your team is allowed to use. It is how you steer people onto the accounts you trust and off the ones you do not.

What you control

• Allow the providers you trust. For example, your enterprise Claude or ChatGPT accounts, which sit under a DPA.
• Warn or block the rest. A personal-account model or an unreviewed provider can be nudged or stopped.
• Applies across surfaces. A blocked model is blocked whether it is reached from the browser, the desktop app, or a coding agent.

Model governance pairs with your category policies: the category rule decides what kind of tool, the model policy decides which underlying model. Together they cover both questions.