Lab 094: Copilot Governance and FinOps: AI Controls, Policies, Budgets, and AI Credits¶
Free-first design
Complete the guided exercise with Copilot Free or the included simulation. Paid features are optional and explicitly labeled.
AI Credit impact
Low for guided work; medium/high for agentic or repeated runs. Prefer automatic model selection, plan before acting, keep context focused, and avoid unnecessary parallel sessions.
What You'll Learn¶
- Map Copilot policies and AI Controls
- Design content exclusion, auditing, and troubleshooting
- Estimate usage in AI Credits without inventing quotas
- Create layered budgets and threshold playbooks
GH-300 alignment¶
Data and architecture (10–15%) · Privacy and safeguards (10–15%). This mapping follows the GH-300 curriculum snapshot dated 2026-08-07. The official outline may retain older product wording such as “Copilot Edits”; this lab teaches the current product experience and calls out that difference.
Concept Map¶
flowchart TD
E[Enterprise credit pool] --> O[Organization budgets]
O --> T[Team alerts]
T --> U[User guidance]
U --> M{Threshold}
M -->|50%| A[Observe]
M -->|75%| B[Optimize context/model]
M -->|90%| C[Investigate outliers]
M -->|100%| D[Apply approved limits]
Hands-on Exercise¶
Step 1: Free simulation: policy matrix¶
For each persona, decide whether completions, chat, CLI, MCP, cloud agent, code review, and third-party agents are allowed. Note that enterprise controls do not govern local VS Code agents in the same way.
Step 2: Free simulation: credit estimator¶
Classify workloads as low, medium, or high. Estimate input tokens × input price + output tokens × output price, then divide USD by 0.01 to express AI Credits. Use current plan docs for real prices.
Step 3: Design layered budgets¶
Allocate an enterprise pool, organization guardrails, team alerts, and user guidance. Define actions at 50%, 75%, 90%, and 100%, including investigation before restriction.
Step 4: Optional paid path¶
In an authorized enterprise, reproduce the matrix in AI Controls, configure audit retention and an MCP allowlist, then compare actual usage with the simulation.
Complete Free Governance Simulation¶
Policy Matrix¶
| Persona | CLI | MCP | Cloud agent | Code review | Decision |
|---|---|---|---|---|---|
| Student | Allow | User-approved local servers | Simulate | Allow if repository enables it | Free-first learning |
| Developer | Allow | Registry allowlist | Selected repositories | Allow | Standard engineering |
| Contractor | Conditional | Block unknown servers | Block | Allow read-only | Restricted data boundary |
| Platform admin | Allow | Manage allowlist | Selected organizations | Allow | Audited administration |
AI Credit Estimator¶
Use this transparent scenario:
| Workload | Input tokens | Output tokens | Input price / 1M | Output price / 1M | Estimated USD | AI Credits |
|---|---|---|---|---|---|---|
| Focused chat | 8,000 | 1,000 | $1.00 | $4.00 | $0.012 | 1.2 |
| Repository plan | 60,000 | 4,000 | $1.00 | $4.00 | $0.076 | 7.6 |
| Agent session | 250,000 | 20,000 | $1.00 | $4.00 | $0.330 | 33.0 |
Formula: USD = input_tokens / 1,000,000 × input_price + output_tokens / 1,000,000 × output_price; AI Credits = USD / 0.01.
These prices are illustrative inputs for the exercise, not a promise of current model pricing. Replace them with the official billing table before a real budget decision.
Supporting kit¶
The site build generates stable PDF and ZIP downloads. The ZIP contains the lab guide and every supporting file declared for this lab.