Purpose
Turn raw account signals into a decision-ready triage summary that tells the CSM or Implementation Manager exactly how worried to be, why, and what to do in the next 48 hours. Eliminates gut-feel account reviews and produces a consistent, defensible risk read across a portfolio.
When to Use
Invoke this skill when the user asks to:
- Assess whether a specific account is at risk of churn, stall, or escalation
- Prepare an account for a portfolio review, QBR prep, or manager 1:1
- Triage an incoming signal — a support spike, dropped usage, a missed milestone, or a stakeholder change
- Decide whether an account needs proactive intervention or can be left alone
- Standardize how risk is called across a book of business
- Draft the "what should I do next" plan for a shaky onboarding
Inputs Required
Before starting, collect:
- Account identity: Account name and current lifecycle stage (onboarding / adoption / renewal / expansion)
- Contract context: Contract value, renewal date, and whether they're on a multi-year or auto-renew term
- Usage/adoption signals: Login frequency, active seats vs. licensed seats, key feature adoption, trend direction (up/flat/down)
- Engagement signals: Last meaningful contact, meeting attendance, email responsiveness, open action items
- Stakeholder status: Is the champion still in place? Any recent departures, reorgs, or new decision-makers?
- Support/sentiment signals: Recent tickets, severity, NPS/CSAT if available, any escalations
- Onboarding/milestone status (if applicable): Time-to-value target vs. actual, milestones hit vs. missed
If usage or sentiment data is missing, score those dimensions as "Unknown" rather than assuming, and flag the data gap in the output.
Instructions
Step 1: Score Risk Across Five Dimensions
Rate each dimension as Healthy (0) / Watch (1) / At Risk (2) / Critical (3). Base each score on the evidence provided, not vibes.
Dimension 1 — Product Adoption & Usage
- Critical: Usage declining 2+ weeks, active seats <40% of licensed, or key feature never adopted
- At Risk: Flat usage below expected benchmark, or adoption stalled at a single use case
- Watch: Usage present but below where a healthy account at this stage should be
- Healthy: Consistent or growing usage, multiple features adopted
Dimension 2 — Relationship & Engagement
- Critical: No meaningful contact in 30+ days, ghosting, or repeated no-shows
- At Risk: Declining responsiveness, meetings postponed, only reactive contact
- Watch: Contact exists but shallow — no forward-looking planning
- Healthy: Regular two-way engagement, proactive from both sides
Dimension 3 — Champion & Stakeholder Stability
- Critical: Champion departed with no replacement, or new decision-maker is skeptical/unknown
- At Risk: Champion role uncertain, reorg in progress, single-threaded relationship
- Watch: One champion, no exec sponsor, thin coverage
- Healthy: Multi-threaded with an engaged exec sponsor
Dimension 4 — Value Realization & Milestones
- Critical: Past time-to-value target with no wins, or milestones missed and slipping
- At Risk: Behind schedule, ROI not yet demonstrated, goals unclear
- Watch: On track but value not yet articulated back to the customer
- Healthy: Milestones hit, measurable outcomes documented
Dimension 5 — Sentiment & Support Health
- Critical: Active escalation, unresolved severity-1 issue, or detractor NPS
- At Risk: Rising ticket volume, unresolved friction, passive/neutral sentiment
- Watch: Occasional friction handled but recurring themes
- Healthy: Low ticket volume, positive or promoter sentiment
Step 2: Assign the Priority Flag
Translate the total score (0–15) and any override rules into a single flag:
- 🔴 RED (Immediate): Total score ≥8, OR any single dimension scored Critical (3), OR renewal within 90 days with total ≥5. Requires action this week.
- 🟡 YELLOW (Monitor): Total score 4–7 with no Critical dimensions. Requires a proactive touch and a plan.
- 🟢 GREEN (Healthy): Total score ≤3. Maintain cadence; look for expansion signals.
Apply override rules explicitly — a champion departure or active escalation forces RED regardless of the arithmetic. State which rule triggered the flag.
Step 3: Identify the Top 3 Risk Factors
Pull the highest-scoring, most consequential drivers. For each:
- What it is: One-line description of the risk
- Evidence: The specific signal from the inputs (e.g., "active seats dropped from 22 to 9 over 3 weeks")
- So what: The consequence if left unaddressed (churn, stalled expansion, escalation)
Step 4: Recommend Immediate Actions
Produce 3–5 concrete actions in priority order. Each action must include:
- Action: A specific, doable step (not "re-engage the customer" — instead "book a 30-min working session with the new ops lead to reset goals")
- Owner: CSM / Implementation Manager / Manager / Exec Sponsor
- Timing: This week / Next 14 days / Before renewal
- Intended effect: Which risk factor it addresses
Step 5: Write the One-Line Verdict
A single sentence a manager can scan: the flag, the core reason, and the single most important next move.
Output Format
- Structure: Header block (account, flag, verdict) → five-dimension scoring table → top 3 risk factors → recommended actions list → data gaps note
- Length: 350–600 words — tight enough to read in under two minutes
- Format: Scannable — a scoring table plus short labeled bullets, not long prose
- Sections: Verdict line → Priority flag → Risk score table (5 dimensions + total) → Top 3 risk factors → Immediate actions (3–5) → Data gaps / assumptions
- Tone: CS operations lead — decisive, evidence-based, no hedging. Call the risk plainly.
Quality Checklist
Before delivering output, verify:
- [ ] All five dimensions are scored with a specific number (0–3) — none left blank; missing data marked "Unknown"
- [ ] The priority flag matches the scoring rules, and any override that forced the flag is named
- [ ] Every risk factor cites a specific signal from the inputs — no generic claims
- [ ] Each recommended action has an owner and a timing window
- [ ] The verdict line states the flag, the reason, and the top next move in one sentence
- [ ] Action count is between 3 and 5, ordered by priority
Notes
- The most common triage error is under-flagging a quiet account — silence in onboarding is a risk signal, not a neutral one. Score "Unknown" engagement toward Watch, not Healthy.
- Renewal proximity is a multiplier: the same score is far more urgent 60 days out than 300 days out. The override rule handles this, but reinforce it in the verdict.
- For portfolio reviews, run this skill per account and stack the verdict lines into a single triage list — RED accounts first.
- A champion departure is the highest-leverage single signal in the model; when detected, the actions should center on re-threading the relationship before anything else.
How to use this skill
This file is a Claude Code skill in Anthropic's standard format — markdown with YAML frontmatter at the top.
Installing in Claude Code:
~/.claude/skills/<folder-name>/SKILL.md— auto-loads in every session anywhere on your machine (recommended for skills you'll use across multiple projects)<your-project>/.claude/skills/<folder-name>/SKILL.md— auto-loads only when working in that project (recommended for project-specific skills)
Restart Claude Code or open a new session to register.
Using in Claude.ai, ChatGPT, Gemini, or other AI tools:
- Drag this file into a new chat, or copy and paste its contents
- Send your task or request as your next message — the AI will apply the skill where it fits
For Claude Code skill details: https://docs.anthropic.com/en/docs/claude-code/skills