Just got admin access for our team’s Cursor plan. After a year of hearing about their “enterprise readiness,” I was expecting at least the basics. The new admin panel is, frankly, a joke for any team that needs actual governance.
Where’s the ability to restrict model usage by cost? No way to set project-level caps or hard limits on AI usage per seat. The audit log is a surface-level activity feed—good luck extracting who prompted what for compliance. The permissions are binary: admin or member. No concept of project-specific roles or even disabling certain features (like the agent) for specific team members. Want to enforce a specific model for a project to control costs? Not happening.
This feels like a feature checklist built for a sales demo, not for a team lead who has to answer to procurement about a runaway bill. They’ve oversold the “Teams” part while delivering a user management page. If you’re evaluating this for a team larger than five, you’re buying a productivity tool with a shared bill, not a managed service.
I’m curious if anyone has found workarounds, or if we’re all just waiting and hoping the admin panel matures before our next budget review.
/charlie
Show me the TCO.
Oof, that's rough. You've hit on a core tension a lot of teams are facing: productivity gains versus governance. That binary permission model is especially painful for larger teams.
You mentioned the audit log being a surface-level feed. That's a huge red flag for any kind of post-incident review or compliance checklist. Without granular logs, you're left blind on both security and cost attribution.
For now, we're in the same boat - logging all API calls through a separate proxy to get the data we need, which defeats the purpose of an integrated "Teams" product. It feels like they built for individual hackers and are now retrofitting teams as an afterthought.
ship early, test often