//OPERATIONS
Live, not staged.
Every number below is read live from the same Postgres tables the pipeline writes to while processing real requests — nothing on this page is mocked or hand-entered. Small numbers are expected: this is a portfolio demo, not a production tenant with thousands of leads/day.
//ALERTS
No stuck workflows
All leads resolved within the 2-minute expected-duration window.
No duplicate writes prevented recently
No replay/duplicate-key collisions in the last 24 hours.
Unsupported-fact rate 0.0% — within target
0/41 drafts had an unsupported claim (target: below 2%).
//BREAKDOWN
Leads by status
- completed64 (100%)
Leads by qualification band
- Sales-ready18 (28%)
- Disqualified16 (25%)
- Needs review13 (20%)
- Nurture10 (16%)
- Insufficient evidence7 (11%)
Identity resolution outcomes
- New (no match)56 (88%)
- Confident match6 (9%)
- Possible match2 (3%)
Outreach draft decisions
- pending_review41 (100%)
//LATENCY PER STAGE (real, from audit_events)
| Stage | Runs | Avg duration |
|---|---|---|
| draft_outreach | 41 | 2.77s |
| extract_facts | 41 | 2.31s |
| verify_facts | 41 | 1.73s |
| classify_message | 64 | 1.72s |
| check_draft_claims | 41 | 1.37s |
| identity_resolution | 64 | 354ms |
| find_research_source | 51 | 157ms |
| propose_crm_change_set | 64 | 154ms |
| qualify | 64 | 151ms |
Scope note
This is the v1 scope from CLAUDE.md section 15: real audit-event data plus 3 alerts (stuck workflows, duplicate-write prevention, unsupported-fact rate). CRM API error tracking, qualification drift over time, and operational controls (pause outbound, replay dead-lettered workflows) are documented in the spec as design, not built — an honest cut line, not a silent shortfall.