Enterprise teams model policy targeting in public preview

Enterprise teams model policy targeting is now in public preview on GitHub. GitHub Enterprise customers with Copilot Business or Copilot Enterprise licenses can use user-based model policy targeting. AI administrators set a baseline model policy and apply it by user rather than only by org-wide defaults.

The feature is policy targeting at the user level inside Copilot on GitHub Enterprise. Administrators define which models are allowed or restricted for specific people under Copilot Business or Copilot Enterprise. That sits on top of existing license-gated Copilot access and gives finer control than a single enterprise-wide model rule.

For engineers and builders, this matters when different teams need different model access under one GitHub Enterprise tenant. Platform and security owners can align model choice with role, risk, or team without spinning up separate orgs. Developers keep working inside the same Copilot surface while admins enforce who may use which models.

Competitively, GitHub is tightening admin control for paid Copilot tiers on Enterprise. User-based model policy targeting is aimed at organizations that already standardize on GitHub and buy Copilot Business or Copilot Enterprise, not at free or individual Copilot users. It treats model access as an identity- and license-scoped control, similar to other enterprise AI admin surfaces that scope capabilities by user or group.

Watch for how baseline policies interact with team and individual overrides once you enable the preview, and confirm which users fall under Copilot Business versus Copilot Enterprise before you rely on targeting. If you run multi-team GitHub Enterprise, plan a small pilot of user-based model policies with a few AI admins and high-sensitivity roles first, then expand after you see how enforcement behaves in day-to-day Copilot use.