Support template
Customer support reply prompt template
On-policy reply to a customer ticket with a fixed structure. Drafting agent-reviewed replies to customer tickets where policy limits (refunds, dates) must never be broken.
The prompt
prompt.md
You are a support specialist for <company>{{company}}</company>.
## Task
Draft a reply to the customer message below.
## Requirements
- Never promise refunds or delivery dates.
- Always include the ticket number <ticket_id>{{ticket_id}}</ticket_id>.
- Keep a friendly, professional tone.
## Output format
Return plain text under 150 words with a greeting, a resolution, and one next step.
## Input
<customer_message>
{{message}}
</customer_message>Variables
| Variable | What to pass in |
|---|---|
{{company}} | Your company or product name, as customers know it. |
{{ticket_id}} | The ticket reference the reply must quote. |
{{message}} | The customer message, exactly as received. |
Each variable sits inside its own tags so the model treats what you insert as data, not instructions.
How it scores
100/100 with no warnings or critical issues, checked in CI on every change.
| Dimension | Score |
|---|---|
| Clarity | 100 |
| Specificity | 100 |
| Completeness | 100 |
| Structure | n/a |
| Consistency | 100 |
| Output specification | 100 |
| Efficiency | 100 |
| Security | 100 |
Tokens and cost per model
Input tokens for the template itself, before you fill in the variables. Costs use reference prices; verify with your provider.
| Model | Tokens | Input cost / 1,000 calls |
|---|---|---|
| GPT-4.1 | 98 | $0.1960 |
| GPT-4.1 mini | 98 | $0.0392 |
| Claude Sonnet 5.5 | 98 (approx.) | $0.1960 |
| Claude Haiku 4.5 | 98 (approx.) | $0.0980 |
| Gemini 2.5 Flash | 98 (approx.) | $0.0294 |
Adapting it
- Add your policy lines as requirements, one per bullet, so each can be checked.
- Change the word limit in the output format rather than adding "be brief", which the engine flags as unquantified.
- Keep the message inside its tags: it is untrusted input and must not be read as instructions.
After editing, the workspace re-scores the prompt as you type and flags anything that regresses.
More templates
- Pull request review: Structured review of a diff with severity-ranked findings.
- Structured data extraction: Extract fields from free text into a strict JSON object.
- Meeting summary: Decisions, owners, and deadlines from a transcript.
- Ticket classification: Route a message to exactly one category with a confidence.
- Tool-using agent system prompt: Scoped agent with explicit tool rules and a stop condition.