churn-analysis
Identify at-risk customer accounts by analyzing usage patterns, engagement signals, and support history to generate churn risk scores and intervention recommendations. Use when the user requests churn analysis or provides relevant inputs for this workflow.
git clone --depth 1 https://github.com/seb1n/awesome-ai-agent-skills /tmp/churn-analysis && cp -r /tmp/churn-analysis/customer-success/churn-analysis ~/.claude/skills/churn-analysisSKILL.md
# Churn Analysis 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. ## Workflow 1. **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). 2. **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. 3. **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. 4. **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. 5. **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. ## Usage 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. ``` ## Examples ### Example 1: Cohort risk analysis with scoring table **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 | | Forge Inc | Enterprise | 22 | Healthy | Stable usage, positive CSAT, expanding seat count | Propose enterprise add-on package, invite to advisory board | | Grid Corp | Pro | 15 | Healthy | Growing feature adoption, 2 referrals made | Send referral program upgrade incentive, case study candidate | | Haven Ltd | Starter | 8 | Healthy | High engagement, recent plan upgrade | Monitor, include in customer spotlight newsletter | **Summary:** 2 accounts critical (25% of ARR at risk), 2 high, 2 medium, 2 healthy. Recommended immediate action on $480K combined ARR in critical tier. ### Example 2: Individual account churn risk deep dive **Input:** "Analyze churn risk for Acme Corp — enterprise account, $240K ARR, 18 days to renewal." **Output:** **Acme Corp — Risk Score: 88/100 (Critical)** | Signal Category | Weight | Score | Evidence | |----------------|--------|-------|----------| | Usage Decline | 35% | 92 | Daily active users dropped from 145 to 41 over 60 days. Core workflow (report generation) usage down 80%. | | Support Sentiment | 25% | 85 | 14 tickets in past 30 days (baseline: 3/month). Two P1 escalations unresolved. CSAT on last 3 tickets: 2, 1, 2. | | Billing Signals | 20% | 78 | Renewal in 18 days, no renewal discussion initiated. Finance team requested contract terms doc (often precedes vendor evaluation). | | Engagement | 20% | 90 | Declined last two QBR invites. Zero email opens in past 30 days. Primary champion (VP Ops) left the company 6 weeks ago. | **Intervention Plan:** 1. **Day 0-2:** VP of CS calls Acme's CTO directly. Acknowledge the degraded experience and unresolved escalations. 2. **Day 2-5:** Assign a dedicated SE to resolve the two open P1 tickets. Provide daily status updates. 3. **Day 5-10:** Schedule an on-site success planning session with the new decision-maker. Present a 90-day recovery roadmap. 4. **Day 10-15:** Offer a concession package: 60-day contr
Design reproducible evaluations for AI agents with representative task sets, explicit rubrics, appropriate graders, baselines, regression gates, and failure analysis. Use when defining agent quality, comparing prompts or models, validating a release, measuring tool-use reliability, investigating regressions, or deciding whether an agent is ready for production.
Design privacy-aware observability for AI agents using traces, spans, structured events, metrics, cost attribution, dashboards, alerts, and investigation workflows. Use when instrumenting an agent, debugging intermittent tool or model failures, defining service-level objectives, analyzing latency or spend, auditing agent decisions, or preparing production monitoring.
Design and verify auditable human oversight, approval gates, escalation paths, and safe state transitions for AI agent workflows. Use when deciding which agent actions require review, adding approve/reject or dual-control flows, preventing unauthorized autonomous effects, creating decision records, reducing rubber-stamping, or recovering safely from rejected, expired, or failed actions.
Design, implement, harden, and verify Model Context Protocol (MCP) servers with precise tool contracts, least-privilege authorization, safe transports, structured errors, and interoperability tests. Use when creating a new MCP server, exposing an API or data source through MCP, reviewing an MCP server design, adding or revising MCP tools, or preparing an MCP server for production.
Design and operate bounded multi-agent workflows with task decomposition, dependency graphs, ownership, handoff contracts, shared-state controls, approvals, recovery, and synthesis. Use when a task contains genuinely independent workstreams, specialized roles, parallel research or implementation, reviewer-worker loops, or coordination problems that one agent should not execute sequentially.
Design and validate model-facing tool definitions with clear names, action-oriented descriptions, bounded JSON Schema parameters, explicit side effects, safe defaults, idempotency, errors, and realistic tests. Use when creating function-calling tools, MCP tools, agent actions, structured tool inputs, or when a model selects the wrong tool, invents arguments, or causes unsafe side effects.
Plan, execute, document, and retest authorized security assessments of AI agents and multi-agent workflows using safe adversarial cases, synthetic identities, canaries, and evidence-based findings. Use when defining red-team rules of engagement, assessing prompt injection or excessive agency, testing tool and identity boundaries, evaluating memory or cross-agent attacks, scoring a campaign, or verifying remediation in an approved environment.
Threat-model and harden AI agents, RAG systems, assistants, and tool-using workflows against direct, indirect, stored, cross-agent, and multimodal prompt injection. Use when reviewing an agent architecture, isolating untrusted content, constraining tools and egress, protecting secrets, adding injection-focused tests, investigating a suspected injection incident, or documenting residual prompt-injection risk.