t0ddc3by/claude-for-customer-success/escalation-memo

/escalation-memo

[PROPOSED]

Open, manage, and close formal escalations with the right stakeholders, the right framing, and a clear resolution path — built from your configured escalation matrix.

Use When

• An account has been confirmed at risk and escalation routing requires a documented memo
• An active escalation needs a status update for leadership or the customer
• An escalation is being closed and a resolution summary is required
• Churn signals have crossed the configured threshold for automatic escalation

Do NOT Use For

• Risk assessment before escalation is confirmed — use /csm:risk-flag first
• Health score review — use /csm:health-score-review
• Renewal commercial prep — use /csm:renewal-readiness
• Routine account updates that don't meet escalation threshold

Typical Activation

"/csm:escalation-memo Acme Corp open" "/csm:escalation-memo Acme Corp update" "/csm:escalation-memo Acme Corp close" "Draft an escalation memo for [account]" "Write up the escalation for [customer]" "Close out the escalation for [account]"

Pre-flight

Read ~/.claude/plugins/config/claude-for-customer-success/csm/CLAUDE.md and ~/.claude/plugins/config/claude-for-customer-success/company-profile.md.

If either is missing or contains [PLACEHOLDER] markers, stop and prompt for /csm:cold-start-interview.

Critical configuration to apply:

• Escalation matrix — routing for each escalation type and severity level
• ARR escalation threshold — determines whether VP CS / CRO are in the loop
• SLA commitments — what response times are promised at each tier
• Escalation owner by type (e.g., Support Manager for P1 technical; VP CS for churn risk)

G-code dependency: All G-code guardrails referenced in this skill (G1–G9) are defined in the CLAUDE.md config loaded above. If Pre-flight halts or config is missing, G-codes are undefined — do not proceed with partial config.

Reasoning Protocol

Full reasoning blueprint: references/reasoning-blueprint.md

Before generating output, apply these primers:

1. CLASSIFY — What type of escalation is this?

• Technical (P1/P2): Unresolved support ticket, SLA breach or approaching breach, product bug with customer impact, engineering involvement required. Route via Support → Engineering path.
• Customer Complaint: Formal dissatisfaction — NPS detractor follow-up unresolved, written complaint, executive frustration, public-facing concern. Emotion is high; separate perception from verified facts.
• Executive Escalation: Named executive at customer requests VP/C-level involvement. Triggered by relationship posture, not health score. Match response seniority to request seniority.
• Internal Process Failure: Customer harmed by internal miss — dropped handoff, broken commitment, internal SLA breach. Name the failure honestly in the internal brief; acknowledge without blame-chaining in the customer draft.

1. CONSTRAINTS — Apply before generating any output:

• G1 — Skill activation: Does this request match escalation-memo? If the issue is a pattern of concern without a triggering event, redirect to /csm:risk-flag.
• G2 — Connector check: Which integrations are needed (Support, CRM, CS Platform, call recording)? Flag any unconfigured or returning stale data.
• G4 — Escalation path required: Verify a named escalation owner is configured in the matrix for this type and severity before generating the memo. No owner = no memo.
• G5 — Confidentiality firewall: Internal language (health scores, escalation IDs, ARR, routing names, revenue-at-risk) never appears in customer-facing drafts. Review customer draft as if the customer's CEO will read it.
• G7 — Revenue language validation: If memo includes ARR, renewal dates, or revenue-at-risk, validate figures with CRM before sharing with leadership or finance.

1. EXPERT CHECK — What a veteran CSM verifies first:

• Is the SLA clock already running? Check status before composing narrative — urgency determines communication depth.
• What has already been promised to the customer? Audit call recordings, ticket notes, and CSM context for "we will" or "by [date]" commitments.
• Is this a repeat escalation type for this account? If CRM shows prior same-type escalations, flag the pattern — this is systemic, not episodic.
• Does ARR exceed the configured threshold? Above-threshold = automatic VP CS / CRO inclusion.
• Route confirmation: Is the escalation owner correct for this type + severity + ARR combination?

1. ANTI-PATTERNS — Domain mistakes to catch before output:

• Citing an SLA window that has already passed without flagging the breach
• Writing recommended actions for the wrong audience (routing not confirmed before drafting)
• Using "I understand your frustration" instead of naming the specific business impact
• Closing an escalation without a root cause hypothesis — even a preliminary one
• Merging internal diagnostic tone with customer-facing communication
• Passive voice in customer drafts to hide internal accountability ("there was a miscommunication" vs. "we missed the handoff")

After execution, verify:

• Does the output match the requested mode (--open/--update/--close) and escalation type?
• Check output against common failure modes for the classified type (see blueprint). Flag any match.
• Is the internal/customer language firewall intact? No internal identifiers in customer draft.
• Confidence: [High] if live integrations confirmed escalation owner, SLA, and account data / [Medium] if partially connected or some data from CSM context / [Low] if user-provided context only — state which.

Mode

--open: Create a new formal escalation record. Produces: escalation brief for the escalation owner, customer-facing acknowledgment draft, and internal tracking block. Default mode.

--update: Update an existing open escalation — new information, status change, resolution progress. Requires the escalation ID or account + type context.

--close: Close an escalation with a resolution summary. Produces a close-out communication for the customer and an internal lessons-learned note.

Escalation type

--type technical: Unresolved P1/P2 ticket; product bug with customer impact; SLA breach. Route: Support → Engineering escalation path.

--type complaint: Customer has formally expressed dissatisfaction — NPS detractor follow-up not resolved, executive complaint, or public-facing concern.

--type executive: An executive stakeholder at the customer has requested VP or C-level involvement. May overlap with churn risk but is triggered by relationship posture, not always by health signals.

--type internal: An internal process failure has affected the customer — missed handoff, broken commitment, SLA breach caused by internal team, not the product.

If no type is specified, ask ONE question to determine escalation type before proceeding.

Data gathering

Connector error categorization: When a connector call fails, distinguish the error type before proceeding:

• Rate-limited (transient): Connector returns HTTP 429 or equivalent throttle signal. Note the rate limit explicitly in output ("CRM data temporarily rate-limited — retry in 60 seconds recommended") and offer to retry rather than proceeding with degraded output.
• Unavailable (permanent for this session): Connector is not configured, authentication has expired, or service is down. Fall back to the manual-input path below and label all affected sections as "connector unavailable — manual input used."

Do not conflate these — a rate-limited connector will return data shortly; an unavailable connector will not.

Pull from connected integrations:

• Support platform: ticket ID, severity, open duration, SLA status, assigned agent
• CRM: ARR, renewal date, contract terms, escalation history for this account
• CS Platform: current health score, lifecycle stage, prior escalations
• Call recording: any recent call where the issue was raised — what was said,

commitments made

If nothing is connected:

"Tell me what's happening. What's the issue? Who's involved? What was the

triggering event? I'll build the escalation record from what you share."

Do not produce a formal escalation without: account name, issue description, the triggering event, and at least one affected stakeholder (internal or customer-facing).

Escalation open structure (--open)

Produce two outputs: Internal escalation brief (full context for the escalation owner) and Customer acknowledgment draft (external, professional).

Internal Escalation Brief

Escalation Brief — [Account Name] [Date] · [Escalation type] escalation · INTERNAL — not for distribution Escalation ID: [ESC-YYYYMMDD-[account abbreviation]] — use this ID in all follow-up

Escalation summary

The issue

What happened — specific, factual, in sequence:

"On [date], [customer contact name] reported [specific issue]. [What was

communicated to the customer at the time — any commitment or ETA given].

The issue has been open for [N] days without resolution. [What has been

tried so far]. The customer is [specific expression of impact or frustration —

quoted if available]."

Do not generalize. Name the specific issue, the specific person who raised it, and the specific business impact the customer described.

Customer impact

Escalation history for this account

Pull from CRM if available. If no prior escalations: "No prior escalations on record for this account."

Pattern note: If this is a repeat escalation type: [review — repeat escalation; root cause may not have been fully resolved]

Internal context (CSM perspective)

What the escalation owner needs to know that is not in the ticket:

• Relationship context: [Is this account at risk? Is the executive sponsor engaged?

Is the champion under pressure?]

• Prior commitments: [Any commitments made by CSM, AE, or Support that are relevant]
• Customer expectation: [What the customer has specifically said they need to

consider this resolved]

• Renewal context: [If renewal is <180 days or account is Yellow/Red: note it]

Escalation routing

Apply the configured escalation matrix exactly.

If ARR meets the configured churn-risk threshold:

"ARR ($[amount]) exceeds the configured escalation threshold ($[threshold]).

Include [configured VP/CRO contact] in escalation communications within [SLA]."

Recommended actions for escalation owner

Specific, not generic. Owner reads this and knows exactly what to do.