GitHub Copilot UBB governance simulator
Interactive AI Credits Flow
Explore how requests are evaluated through user-level controls, shared credits, and metered guardrails. Use single-request and multi-step agentic flows to demonstrate where requests pass, warn, or stop.
User used
0
Pool
0
Cost centre
0
Enterprise cap
0
Controls
Change guardrails and re-run the simulation.
Default hard cap per user.
Used for architect/power users.
Included credits before paid mode.
Checked after pool exhaustion.
Final global cap after pool exhaustion.
Simulation mode
Run a single request or a multi-step agentic workflow.
Example: 1,200 for normal usage; 4,000 to exceed a 2,500 ULB.
Flow diagram
1. User-level control
2. Shared pool
3. Paid mode
Only after pool is exhausted.
4. Cost centre
5. Enterprise cap
Result
Run a scenario.
Execution log
Run a scenario to see the decision trail.