Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →seb1n/churn-analysis
Identify at-risk customer accounts by analyzing usage patterns, engagement signals, and support history to generate churn risk scores and intervention recommendations.
Passed 16 of 22 test cases with the skill, and 9 of 22 without it.
and 8 more in the full report
7 fixed9 passed either way6 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
We conducted 15 exit interview calls with accounts that canceled their subscriptions last month. Here are the transcript summaries. Categorize the root causes of cancellation into primary categories like product missing features, price sensitivity, or internal company restructuring, and calculate the percentage distribution across categories.
| Checked | Without skill | With skill |
|---|---|---|
| Categorizes cancellation root causes from exit interviews. | Missed | Passed |
| Does not score active accounts for churn risk. | Missed | Passed |
The instructions your model reads when the skill loads.
Detect early warning signs of customer churn by aggregating usage telemetry, support interactions, billing history, and engagement metrics into a composite risk score per account. This skill segments accounts into risk tiers and produces actionable intervention playbooks tailored to each tier, enabling CS teams to proactively retain revenue.
Collect usage and engagement data — Pull metrics across product analytics (DAU, feature adoption, session duration), support history (ticket volume, CSAT scores, escalations), billing signals (late payments, downgrade requests, contract end dates), and engagement touchpoints (email opens, webinar attendance, QBR participation). Normalize all metrics to a consistent time window (typically 90 days trailing).
Define churn signals — Establish the leading indicators that correlate with churn in your specific context. Common signals include: login frequency dropping below 50% of the account's historical average, a spike in support tickets (3x baseline) within 30 days, missed or late renewal payment, champion contact leaving the company, feature adoption plateau (no new features used in 60 days), and declining NPS scores on consecutive surveys.
Score risk per account — Compute a weighted composite score from 0 (healthy) to 100 (imminent churn) for each account. Weight signals by their predictive power — usage decline typically carries 35% weight, support sentiment 25%, billing signals 20%, and engagement metrics 20%. Adjust weights based on historical churn correlation data if available. Accounts missing data for a signal category receive a neutral score for that dimension with a data-quality flag.
Segment into risk tiers — Bucket accounts into four tiers based on their composite score: Critical (75-100) — immediate intervention required, likely to churn within 30 days. High (50-74) — concerning trends, intervention needed within 2 weeks. Medium (25-49) — early warning signs, monitor and engage proactively. Healthy (0-24) — on track, maintain regular cadence.
Generate intervention recommendations — For each tier, produce specific action plans. Critical: executive sponsor outreach, emergency success plan, potential concessions or credits. High: CSM-led deep dive call, custom training session, product roadmap preview. Medium: automated check-in email sequence, in-app tips targeting underused features, invite to community events. Healthy: upsell/cross-sell opportunity identification, referral program invitation.
Provide account data or describe the account portfolio you want analyzed. The agent will compute risk scores and return tiered recommendations.
Analyze the churn risk for our Q1 cohort of 200 accounts. Here's the usage
data export. Identify the top 10 at-risk accounts and recommend interventions.
Input: Usage and engagement data for 8 accounts over the past 90 days.
Output:
| Account | Plan | Risk Score | Tier | Key Signals | Recommended Action |
|---|---|---|---|---|---|
| Acme Corp | Enterprise | 88 | Critical | Logins down 72%, 14 tickets in 30d, renewal in 18d | Exec sponsor call within 48h, offer dedicated onboarding reset, prepare 2-month extension |
| Bolt Inc | Pro | 71 | High | Feature adoption stalled, champion left org, NPS dropped from 8→4 | CSM deep-dive on use cases, identify new champion, schedule product roadmap session |
| Cedar Ltd | Enterprise | 63 | High | API usage down 45%, 2 escalated tickets, QBR declined | Technical health check, assign SE for integration review, CSM outreach to new stakeholder |
| Dash Co | Pro | 42 | Medium | Session duration declining, stopped attending webinars | Trigger re-engagement email series, in-app walkthrough for new features launched in Q4 |
| Echo LLC | Starter | 38 | Medium | Login frequency down 30%, no support contact in 60d | Automated check-in email, offer free training session |
test cases passed, vs 9 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.