Build a copy-ready customer support reply prompt with issue facts, policy limits, empathy tone, resolution options, escalation checks, and privacy safeguards.
Use this tool when you need an AI assistant to draft a customer support reply that is helpful without overpromising. Fill in the customer issue, verified facts, policy constraints, desired resolution, and tone; then copy a prompt that asks the AI to separate empathy, action steps, escalation risks, and claims that need human review.
Use this structure before asking AI for a draft: reply quality = verified facts + policy boundaries + clear next step + empathy + escalation trigger. If one part is weak, the generated prompt tells the AI to identify missing information instead of fabricating certainty.
| Section | What the AI should produce | Quality check |
|---|---|---|
| Opening | Acknowledge the issue and emotion without admitting unverified fault. | Would a frustrated customer feel heard in the first two sentences? |
| Facts | List only confirmed account, timing, plan, or policy facts. | Are private details removed and assumptions labeled? |
| Resolution | Give the next step, owner, expected timing, and any customer action. | Can the customer act without asking a follow-up question? |
| Escalation | Flag refund, legal, billing, identity, or safety cases for human review. | Does the reply avoid promises a support agent cannot keep? |
Red Mode rescue on 2026-05-17: existing URL upgraded from a thin static prompt into an original browser-side prompt builder with readiness scoring, task-specific fields, FAQ, JSON-LD, and stronger internal linking. No external repository content copied.