copilot-metrics-analyzerlisted

# copilot-metrics-analyzer A Category 3 ops tool. The org's Copilot Enterprise tenant produces usage metrics; this skill translates raw API output into the answers finance / ops / managers actually ask: "who's expensive, who's idle, are we within budget, what's the chargeback per BU?" ## When to use - Weekly cost monitoring during a rollout (Phase 3 / Phase 5 of [`copilot-rollout-playbook`](../copilot-rollout-playbook/SKILL.md)). - Cost-attribution conversation with finance — "show me the per-LOB Copilot spend with the org's actual surcharge applied." - Identifying dormant licenses for cost recovery. - Outlier detection — anomalously high acceptance rate or AI-credits consumption that warrants 1:1 conversation. ## When NOT to use - Strategic Copilot adoption decisions — those are research, not metrics. - Performance / latency monitoring — that's a different signal entirely. - Individual developer evaluation — these metrics are aggregate, not per-developer-quality. ## Input The Copilot Usage Metrics API JSON, per-user daily granularity. Required fields (per [[usage-metrics-ai-credits-per-user]]): ```json [ { "user": "dev1@example.com", "date": "2026-06-19", "active": true, "suggestions_shown": 142, "suggestions_accepted": 38, "chat_turns": 12, "ai_credits_used": 18.4 } ] ``` Plus optional context the user provides: - `period_start` / `period_end` — analysis window - `business_units` — mapping of user → BU for chargeback - `cost_per_cred