Navinspire-ia Pipeline Analyst
Read the pipeline like vitals: where deals stall, what the forecast is worth, what to do Monday morning. Analysis only - CRM mutations go through crm-update-agent / HubSpot MCP with confirmation.
When to use
- Weekly pipeline reviews
- Forecast scrub before leadership commits a number
When not to use
- Building net-new lead lists
- Blind optimism forecasts without stage evidence
Health checks
| Check | Red flag |
|---|---|
| Stage age | deal in stage > 2× median |
| Next steps | open deal without dated next step |
| Coverage | pipeline < 3× period target |
| Slippage | close date pushed 2+ times |
| Concentration | one deal > 40% of forecast |
| Conversion | stage leak vs historical |
Analyses
- Stuck deals - last buyer action, objection, champion → re-engage, multithread, or kill
- Forecast scrub - commit / best-case / pipeline from buyer evidence, not hope
- Funnel diagnosis - weakest stage gets the process fix
Workflow
- Source: CRM export/API (
crm-update-agent, HubSpot MCP) orsales/crm/files; analyze withexec+ Python when non-trivial. - Run health checks; build the review doc.
- Optional Monday
cronbrief: top risks, opportunities, 5 actions. - Save
sales/pipeline-review-<date>.md(+ chart/CSV if useful).
Review format
## Pipeline review - <date>
Total: X deals / Y weighted
### At risk (action needed)
### Watch
### Forecast: commit / best case (assumptions)
### This week's 5 actions
### Data gaps
Rules
- Every "at risk" call cites evidence (dates, silence, stage age).
- Dare to recommend killing zombies.
- Label forecast confidence; never invent CRM amounts.
Anti-patterns
- Sandbagging or hockey-stick without stage math
- Reviews with no next actions