# Create an escalation and human-handoff guide

When and how to escalate issues, or hand off from a bot to a human.

- Topic: Customer Support
- Framework: RCTFVV
- Level: Advanced
- Tags: Customer Support, Fill-in Template
- Source: https://www.thinkwithprompt.com/prompts/create-an-escalation-and-human-handoff-guide

## Prompt

```text
Role: Act as a support team lead defining escalation and bot-to-human handoff rules.

Context: Channels and any automated/bot tier: [CHANNELS_BOT]. Escalation tiers and specialisms: [TIERS]. Situations that should always escalate (risk, VIP, legal, safety, repeated failure, high emotion): [ALWAYS_ESCALATE]. Situations that should stay in tier 1: [STAY_TIER1]. SLA and after-hours rules: [SLA_HOURS]. What context systems can pass: [CONTEXT_AVAILABLE].

Task: Produce: (1) the escalation triggers, each with the signal that identifies it and the destination tier/team; (2) the bot-to-human handoff triggers and how the handoff is done (what the customer sees, what context is passed, warm vs cold); (3) an anti-pattern list - when NOT to escalate; (4) the escalation note template (summary, what's been tried, why escalating, customer state, desired outcome); (5) 3 worked examples.

Format: Sections: Escalation triggers (table) | Handoff triggers & mechanics | Do-not-escalate | Escalation note template | Examples.

Verification: Ensure every 'always escalate' situation has a trigger and a destination; flag any trigger that relies on a signal the systems cannot actually detect.

Validation: Run a high-emotion but low-severity contact and a calm but high-risk one through the guide: does each go to the right place? Adjust the triggers that misfire.
```

## What to fill in

- CHANNELS_BOT
- TIERS
- ALWAYS_ESCALATE
- STAY_TIER1
- SLA_HOURS
- CONTEXT_AVAILABLE

Free to copy and use. From ThinkWithPrompt (https://www.thinkwithprompt.com/), a library of 2,663 curated AI prompts.