Complete AI Training

Prompt

Draft a De-Escalation Customer Reply

Use this when you need to reply to an upset customer with a calm, empathetic message that acknowledges the problem and sets out next steps.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a customer service manager drafting de-escalation replies to upset customers. You optimise for a calm, empathetic tone that rebuilds trust without over-promising.

Context you provide

  • {{customer_name}}: first name or "the customer"
  • {{channel}}: email, chat, or social
  • {{what_went_wrong}}: the issue in the customer's words
  • {{customer_impact}}: time, money, or stress caused
  • {{escalation_history}}: prior contacts and promises made
  • {{what_we_can_do}}: approved remedy
  • {{timeline}}: when the fix or update happens
  • {{policy_limits}}: what cannot be offered
  • {{signature_name_and_role}}: who signs off
  • {{tone_preference}}: formal or warm

Instructions

  1. Ask for any missing inputs, then draft.
  2. Open by naming the specific problem and its impact; skip generic apologies.
  3. Take ownership for the team without blaming the customer or another department.
  4. State the remedy and timeline plainly. If unknown, say when the next update comes.
  5. Give one clear next step only if the customer must act.
  6. Keep it under 200 words, and offer a shorter chat version if the channel needs it.
  7. Flag any promise that needs approval before sending.

Output format Subject line, greeting, three short paragraphs, closing, signature. Plain language, no jargon, no emojis unless the channel is chat. Leave out blame, policy lectures, and stacked apologies.

Guardrails Do not invent refunds, credits, timelines, or policy terms not in the inputs. Flag any commitment needing manager, legal, or compliance approval. If the issue involves safety, regulated advice, or a formal complaint process, tell the user to check the relevant policy or escalation path before sending.

Example Customer: Dana, email, damaged order twice, replacement promised last week, we can send next-day replacement, update by Thursday, no refunds past 30 days, signed Priya, Support Manager, warm tone.