Example prompts
What this Skill does
Scores every account on four behavioural risk signals over the last 90 days: current under-usage, activity decline, a cancel or intent event, and single-user adoption. Accounts tripping at least 2 (or 3) of them form the cohort. One weak signal is noise; several at once rarely is.
When it fires
- At least 2 (or 3) of 4 risk signals in the last 90 days
- Signals: under-usage, decline, intent event, single active user
- Highest priority when the renewal is due within 90 days
Inputs
- Brazebee MCP: connected to Claude or ChatGPT with your Brazebee API key
- Product usage: Amplitude, PostHog, or Mixpanel connected in Brazebee
- Billing and CRM: Stripe and HubSpot for ARR, plan, renewal date, and account owner
How it works
- 01 / Runs the four individual risk checks over the last 90 days
- 02 / Counts how many each account trips
- 03 / Builds cohorts for at least 2 and at least 3 signals
- 04 / Compares their churn rate and ranks accounts by ARR
Output
Churn Intent: at least 2 of 4 risk signals (last 90 days)
Cohort: 58 accounts · $520k ARR
Churn: 27% in cohort vs 3% in the rest (9.0x)
Act on: 14 accounts with renewal in the next 90 days · $160k ARR
Route: CSM lead · Slack #cs-alerts · "High-priority save plan" Example numbers. Your thresholds and rates come from your own data.
Tips
- Start here if you only switch on one churn Skill: it has the fewest false alarms
- Show which signals tripped in the brief, so the CSM knows what to fix
- Sentiment from calls and emails would add a fifth signal once that data is connected
