Ticket Triage

Every support inbox drowns in the same problem: a billing question filed as a bug, an angry customer marked low priority, and a one-line summary that says nothing. This prompt triages a support ticket into a strict shape — category (billing, technical, or sales), an urgency score from 1 to 5, sentiment, and a single-sentence summary under 25 words. The enums keep your routing rules deterministic: no free-text category that breaks your filters, no urgency of 9 on a 1-to-5 scale. The classic traps are confusing urgency with sentiment (an angry customer is not automatically a critical outage) and letting the model invent details the customer never wrote. Use it in front of your helpdesk, your Slack triage channel, or an auto-responder that must decide in under a second.

Optimized System Prompt

You are a support ticket triage engine. Read the ticket and return ONLY a valid JSON object matching the schema below. category must be exactly one of: billing, technical, sales. urgency is an integer from 1 (low) to 5 (critical). sentiment must be exactly one of: positive, neutral, negative. summary is a single sentence, max 25 words, in the ticket's language. Do not add explanations, markdown, or extra fields. Never invent details that are not in the ticket.

Raw JSON Schema

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "SupportTicket",
  "type": "object",
  "additionalProperties": false,
  "required": [
    "category",
    "urgency",
    "sentiment",
    "summary"
  ],
  "properties": {
    "category": {
      "type": "string",
      "enum": [
        "billing",
        "technical",
        "sales"
      ]
    },
    "urgency": {
      "type": "integer",
      "minimum": 1,
      "maximum": 5,
      "description": "1 = low, 5 = critical."
    },
    "sentiment": {
      "type": "string",
      "enum": [
        "positive",
        "neutral",
        "negative"
      ]
    },
    "summary": {
      "type": "string",
      "minLength": 1,
      "maxLength": 200,
      "description": "One sentence, max 25 words."
    }
  }
}

Validation Code — Zod (TypeScript)

import { z } from "zod";

export const TicketSchema = z.object({
  category: z.enum(["billing", "technical", "sales"]),
  urgency: z.number().int().min(1).max(5),
  sentiment: z.enum(["positive", "neutral", "negative"]),
  summary: z.string().min(1).max(200),
});

export type Ticket = z.infer<typeof TicketSchema>;

Validation Code — Pydantic v2 (Python)

from pydantic import BaseModel, Field
from typing import Literal


class Ticket(BaseModel):
    category: Literal["billing", "technical", "sales"]
    urgency: int = Field(..., ge=1, le=5)
    sentiment: Literal["positive", "neutral", "negative"]
    summary: str = Field(..., min_length=1, max_length=200)

Try it in the StrictJSON tool → · ← Back to the Library