Deal Stage Auditor
Catch deals stuck in the wrong stage before your forecast breaks
Deal stages lie because reps don't update them. This skill analyzes email activity, meeting patterns, and conversation signals to flag deals whose behavior doesn't match their reported stage.
INGREDIENTS
PROMPT
Create a skill called "Deal Stage Auditor". Analyze all open deals in my pipeline from my CRM or a current pipeline export. For each deal, compare the reported stage against actual activity signals: email frequency, meeting recency, document sharing, stakeholder engagement count, and days in stage. Flag mismatches where behavior doesn't match the stage (e.g., deal in "Negotiation" with no activity in 14+ days, or deal in "Discovery" where pricing docs were shared). For each flagged deal, show: current stage, suggested stage, evidence for the suggestion, and days since last meaningful activity. Generate a weekly audit summary grouped by rep.
How It Works
The skill cross-references what a deal's stage says against what the actual activity
signals show. A deal marked "Negotiation" with no email activity in 3 weeks gets
flagged. A deal in "Discovery" where pricing was discussed gets nudged forward.
What You Get
- Stage-vs-activity mismatch detection across all open deals
- Suggested stage corrections with supporting evidence
- Inactivity alerts by stage (configurable thresholds)
- Weekly audit report for pipeline review prep
- Historical accuracy tracking: which reps' stages are most reliable?
Setup Steps
- Define your sales stages and expected activity patterns per stage
- Set inactivity thresholds per stage (e.g., Discovery = 7 days, Negotiation = 3 days)
- Connect email and calendar data, and provide pipeline data or a recurring CRM export for stage analysis
- Run the first audit manually to calibrate
Tips
- Don't auto-change stages — flag and let reps confirm. Trust but verify.
- The weekly audit report is your best pipeline review prep tool
- Track rep-level accuracy over time for coaching conversations
- Pair with the Forecast Modeler for stage-corrected forecasts