Observe
Gain total visibility into all AI traffic flowing through your organization. Every query, every model, every dollar — tracked in real time.
Visibility Features
- Dashboards — Real-time dashboards showing query volume, governance actions, cost trends, and orchestration metrics.
- Query Log — Every query logged with full context: original query, improved query, response, governance actions, and compression stats.
- Cost Controls — Track AI spend by model, user, department, and routing rule. See savings from caching, routing, and compression.
- Metrics & Analytics — Comprehensive metrics including context hit rate, tool call counts, workflow executions, and compression savings.
What Xilos Tracks
For every query that passes through the platform, Xilos records:
- Original query — The exact user input
- Improved query — The processed query after enrichment and compression
- Response — The LLM's response
- Routing decision — Which model was selected and why
- Governance actions — Any restrictions, blocks, masks, or flags applied
- Compression stats — Token count before and after compression
- Cost breakdown — Input tokens, output tokens, and total cost
- Quality scores — Faithfulness, answer relevance, context relevance
- Cache hit/miss — Whether the response was served from cache
- Latency — Time to first token and total response time
- User and department — Who made the query
Real-Time Monitoring
The dashboard updates in real time, showing:
- Query volume over time (activity chart)
- Query treatment breakdown (unrestricted, blocked, flagged, masked)
- Trending topics in unrestricted and blocked queries
- Executive report with AI-generated insights
Historical Analysis
For any historical date range, Xilos provides:
- Period-over-period comparison
- Donut charts for query treatment breakdown
- Executive report with contextual insights
- Drill-down capability into specific categories
- Export for custom reporting