Skip to content This library is free, offered as dana. If it helps your work, donate here →
penguin tree ai
← back to libraryDonate here

This is one of 3,622 free prompts in the penguin tree ai library. Sign up once and the whole library opens, with new prompts by email.

open the library →

Case escalation decision tool

Analyzes a support case against a structured escalation decision framework and returns a clear escalate-or-resolve verdict, the correct escalation tier, a ready-to-send handoff summary, and the resolution path if no…

download .md

Purpose

Turn an ambiguous support case into a defensible escalation decision — telling the agent whether to resolve it themselves or escalate, which tier to escalate to, and exactly what to include in the handoff — so cases route correctly the first time and nothing bounces between queues.

When to Use

Invoke this skill when the user asks to:

  • Decide whether a specific case should be escalated or handled at the current level
  • Determine which team or tier a case should go to (Tier 2, engineering, billing, account management, legal/trust & safety)
  • Write an escalation handoff summary for a case being passed on
  • Check a case against escalation policy before deviating from standard workflow
  • Handle an angry, at-risk, or high-value customer and are unsure of the right path
  • Assess whether a recurring or edge-case issue exceeds their authority to resolve

Inputs Required

Before starting, collect:

  • Case summary: What the customer is reporting or requesting, in 2–4 sentences
  • What's been tried: Steps the agent has already attempted and the results
  • Customer signals: Sentiment, tenure/value if known, any threats to churn or escalate externally
  • Blocker type: Why the agent is stuck — missing access/permissions, unknown root cause, policy limit, technical bug, financial impact, or legal/safety concern
  • Available tiers/teams: Who the agent can escalate to in their org (defaults to Tier 2, Engineering, Billing, Account Management, and Trust & Safety if not specified)
  • SLA / time context: How long the case has been open and any deadline
  • Agent authority limits: Any known caps on refunds, credits, or account actions the agent can take

If authority limits or available tiers aren't provided, use standard defaults and flag every decision that depends on an assumption the agent should confirm.

Instructions

Step 1: Run the Escalation Trigger Check

Evaluate the case against six escalation triggers. Any single YES on a hard trigger means escalate; soft triggers require judgment and are weighed together.

Hard triggers (any one = escalate):

  • Authority breach: Resolution requires an action beyond the agent's stated limits (refund above cap, account deletion, contract change, pricing exception)
  • Access/permission gap: Fixing the issue requires system access or tooling the agent does not have
  • Legal / safety / compliance signal: Threats, harm, fraud, data privacy, regulatory, or trust & safety concerns
  • Confirmed defect: A reproducible bug or platform fault that no configuration change can resolve

Soft triggers (weigh together):

  • Repeated failure: Case has already bounced or the agent has exhausted the known playbook with no resolution
  • Customer risk: High-value, at-risk, or publicly vocal customer whose sentiment is deteriorating

State which triggers fired and quote the input that fired them.

Step 2: Apply the Workflow-Deviation Test

If no hard trigger fired, determine whether resolving independently would deviate from standard workflow.

  • Within workflow → resolve independently; the agent has both the authority and the path
  • Deviation with precedent → resolve independently but document the deviation and reasoning
  • Deviation without precedent → escalate for approval before acting, even if the agent technically could act

The default bias: if resolving requires an undocumented exception, escalate for approval rather than setting an unofficial precedent.

Step 3: Select the Escalation Tier (if escalating)

Map the primary blocker to the correct destination:

  • Tier 2 / Senior support: Complex but known issue type, exhausted playbook, no defect or authority gap
  • Engineering / Product: Confirmed or suspected defect, needs code/config investigation
  • Billing / Finance: Refund/credit above cap, disputed charges, subscription/contract mechanics
  • Account Management / CSM: Relationship risk, churn threat, high-value renewal, commercial ask
  • Trust & Safety / Legal: Harm, fraud, privacy, regulatory, or threat signals

If multiple destinations apply, name the primary owner and list secondary teams to CC.

Step 4: Set Urgency

Assign a priority based on impact × time sensitivity:

  • P1 – Immediate: Safety/legal, outage-level defect, or imminent churn of a high-value account
  • P2 – Same day: Blocked customer, SLA at risk, deteriorating sentiment
  • P3 – Standard queue: Non-blocking, no SLA pressure

Step 5: Produce the Verdict and Artifact

  • If resolve independently: give the specific resolution path, the exact actions the agent is authorized to take, and any documentation note to log.
  • If escalate: produce a copy-paste handoff summary (see Output Format) addressed to the selected tier.

Output Format

  • Structure: A verdict line, the trigger evidence, tier + urgency (if escalating), and either a resolution path or a handoff summary
  • Length: 250–450 words
  • Format: Structured document with labeled sections; the handoff summary in a clearly delimited, copy-paste block
  • Sections: Verdict → Triggers Fired → Workflow-Deviation Test → Tier & Urgency (if applicable) → Resolution Path OR Handoff Summary → Confirm-Before-Acting flags
  • Tone: Decisive and operational — a team lead giving a clear call, not hedged options

The Handoff Summary block must contain: (1) one-line issue statement, (2) customer/account and value signal, (3) steps already tried and results, (4) exact reason for escalation, (5) specific ask of the receiving team, (6) priority.

Quality Checklist

Before delivering output, verify:

  • [ ] A single clear verdict is stated — escalate or resolve, no "it depends" ending
  • [ ] Every fired trigger cites the specific input that triggered it
  • [ ] If escalating, exactly one primary tier is named (secondary teams marked as CC)
  • [ ] A P1/P2/P3 urgency level is assigned with a one-line reason
  • [ ] The handoff summary contains all six required elements and is copy-paste ready
  • [ ] Any decision resting on an assumed authority limit or tier is flagged for the agent to confirm

Notes

  • The most common routing error is escalating an authority problem to Engineering — an authority gap goes to whoever owns the approval, not whoever owns the system.
  • A case that has already bounced once should escalate up a tier rather than sideways; lateral transfers are where cases die.
  • When a hard trigger and "resolve independently" seem to conflict, the hard trigger always wins — undocumented exceptions create precedent risk that outlasts the single case.

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:

  1. Drag this file into a new chat, or copy and paste its contents
  2. 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