If you run a Claude Enterprise org with a few hundred people in it, you know the problem: everything goes through the console, and everything is manual. As of July 14 there’s an alternative. The Admin API now does user management — in beta, for all Claude Enterprise organizations.
What you get
- List members and look them up by email address
- Change a member’s role
- Remove members
- Send and withdraw invites
- Manage groups and their membership
- Read custom roles
The catch is in the details: group and custom-role requests need the beta header anthropic-beta: ce-user-management-2026-07-13. Member and invite requests take no header. And an Admin API key with the read:org_audit scope can call every user-management GET endpoint — handy for audit tooling that should only ever read.
Why it matters
This is the unglamorous but decisive part of enterprise software: onboarding and offboarding have to fit into processes that already exist. When someone leaves the company, nobody should have to remember to log into the Claude console. A script should handle it — ideally the same one that tears down the Okta account.
Put this next to June’s enterprise-managed auth for MCP connectors and July’s spend controls, and a picture forms: Anthropic is building the administrative layer IT departments need so they don’t have to treat Claude as a special case.
My take
Honestly, I find releases like this more interesting than the next model. A better model does nothing for you if your IT can’t roll the tool out without inventing a bespoke process around it. This is where it gets decided whether AI in a company stays a pilot or becomes infrastructure.
It’s still beta, though. If you automate on top of it, keep that beta header in mind — names with dates baked into them have a habit of getting replaced.
Sources: