Script to calm an angry customer and escalate without making them repeat

A support script to de-escalate an angry customer by phone or chat, offer options within your authority and escalate to tier 2 with a warm handoff so they never have to repeat themselves. Designed for ChatGPT, Claude, Gemini and Microsoft Copilot. It includes 4 customizable variables and you can copy it for free in one click.

Intermediate · Text · Pending editorial review

Prompt

Act as a customer service trainer who specializes in conflict de-escalation and escalation handling.

Context:
- Typical situation causing the anger: {{situation}}
- Channel: {{channel}}
- What the agent can offer without approval: {{authority}}
- Who it escalates to and their response time: {{tier_two}}

Deliverables:
1. A phase-by-phase script with word-for-word lines adapted to the {{channel}} channel:
   a) Opening and listening: let the customer explain without interrupting, and show you listened (summarize in their words).
   b) Acknowledgment: 5 lines that validate the emotion without admitting fault that has not been established.
   c) Diagnosis: 4-6 questions to understand the case and verify details.
   d) Options: how to present 2-3 choices within {{authority}} so the customer picks.
   e) Escalation: clear criteria for escalating and a line that explains it without sounding like a brush-off.
   f) Warm handoff: what to tell the customer and what to tell {{tier_two}}, so the customer does not repeat anything.
   g) Close: recap of what was agreed, next step, a realistic timeframe and how to get back in touch.
2. Boundaries for insults or threats: a respectful warning, a second warning and how to end the conversation under company policy.
3. Ten phrases that make things worse, each with an alternative.
4. An internal handoff note template (issue, what was tried, customer expectation, urgency).
5. One full practice dialogue for training (role-play).

Constraints:
- The agent never promises anything beyond their authority or timelines they do not control. Where a policy is missing, use [FILL IN].
- Human tone, no scripted lines repeated on a loop.
- If key information is missing, ask me first.

Format: script as a table (phase, goal, sample lines), followed by sections 2-5 under headings.

How to use it

  1. Define the agent’s real authority precisely (amount, type of compensation). 2. Run the role-play in pairs at your next team meeting. 3. Review the handoff template with tier 2 so it includes everything they need.

Usage example

Situation: third failed sofa delivery · Channel: phone · Authority: refund shipping or a $30 credit · Tier 2: logistics, replies within 24 hours.

Sample output

a) Opening and listening

"Hi, this is [AGENT]. Tell me what happened, I'm listening." Let them talk without interrupting, then summarize in their words: "Let me make sure I've got this: you bought on [DATE] and your card shows two charges for the same amount. Is that right?"

b) Acknowledgment

  • "I completely understand why an extra charge is worrying."
  • "It makes sense that you want this sorted today."
  • "Thanks for flagging it. Let's look at it together."

c) Diagnosis

  1. Can you confirm your order number?
  2. Are both charges showing as posted, or is one pending?
  3. What date and amount does each one show?

d) Options (within your authority)

"If I confirm it's a duplicate, I can refund the extra charge right now and, for the hassle, add a credit of up to $15. Does that work for you?"

e) Escalation

If the charge isn't in our system or the customer asks for more than you can offer: "For this I need our payments team, who can see the payment gateway. You'll hear back within 2 business days, and here's your case number: [CASE #]."

Phrases that make it worse

AvoidInstead
"Calm down""I understand this is frustrating. Let's fix it."
"That's your bank's problem""Let me check what happened on our side."

…

Key facts

Category
Customer service · Support scripts
Output type
Text
Level
Intermediate
Designed for
ChatGPT, Claude, Gemini and Microsoft Copilot
Verified in
Pending editorial verification
Variables
Situation, Channel, Agent authority and Tier 2 and response time
Language
English
Price
Free, no sign-up
Updated
Nov 7, 2026

Frequently asked questions

Is acknowledging the customer’s emotion the same as admitting fault?

No. Saying "I understand it is frustrating to wait three weeks" validates the experience without conceding an error that has not been confirmed. The script keeps the two apart.

What is a warm handoff?

The agent briefs the tier 2 person before transferring the customer, or leaves a complete note. That way the customer never has to tell their story again.

What is this prompt for?

A support script to de-escalate an angry customer by phone or chat, offer options within your authority and escalate to tier 2 with a warm handoff so they never have to repeat themselves.

Which AI tools does it work with?

It is designed for ChatGPT, Claude, Gemini and Microsoft Copilot. We have not verified it in these tools yet, so results may vary between them.

How do I customize the prompt?

Click “Customize” and fill in the variables (Situation, Channel, Agent authority and Tier 2 and response time). The text updates instantly and you can then copy it in one click. Values are stored only in your browser.

Do I need experience to use it?

It is intermediate: it helps to know the topic so you can provide good context and review the result critically.

Is this prompt free?

Yes. Every HomeCasa Prompts prompt is free, with no sign-up or subscription. You can also use them in professional projects.

Can I trust the AI output?

Treat it as a high-quality draft: always check data, figures and important claims before publishing or making decisions.

Related prompts