Command Center
Operational health, usage, events, traceability and billing in one cockpit
Executive picture complete
Revenue, operations, billing risks and AI recommendations are ready for action.
Actions
Filter the whole command center and export what you see
AI recommendations
Generated by the FinOps analysis engine from live usage, margin and utilisation signals
Move 38% of support traffic to nemotron-ultra-550b
This monthSupport Copilot runs 84% of prompts under 600 tokens on gpt-4o. Quality scores stay within 2% on the value tier.
Monthly impact
$41,200
Confidence
92%
Renegotiate Helios AI before renewal
Next 30 daysHelios AI is at 31% gross margin with a declining growth rate and an overdue invoice — 9 points below portfolio average.
Monthly impact
$22,600
Confidence
78%
Reclaim 12 idle GPUs in cluster us-east-2
This weekAverage utilisation below 20% for 7 consecutive days while queue depth elsewhere stays near zero. Consolidation is safe.
Monthly impact
$55,260
Confidence
88%
Raise Northwind's committed token floor
Next quarterNorthwind has run 127% of contracted volume for 3 months. A higher commit converts overage into predictable ARR.
Monthly impact
$118,000
Confidence
71%
Cluster health
Four clusters · 8 × B300 each
Cluster-A
Healthy94%
Cluster-B
Healthy91%
Cluster-C
Healthy93%
Cluster-D
Healthy90%
Model health
Latency and error rate per model
| Model | p95 | Err % | Status |
|---|---|---|---|
| nemotron-ultra-550b | 412 ms | 0.02% | Healthy |
| gpt-4o | 388 ms | 0.05% | Healthy |
| claude-3.5-sonnet | 964 ms | 1.27% | Degraded |
| llama-3.1-70b | 275 ms | 0.01% | Healthy |
| qwen-2.5-72b | 231 ms | 0.03% | Healthy |