Build a Support Ticket Triage Agent with Oxlo.ai

Описание и применение

Создайте агента для первичной сортировки заявок в службу поддержки, используя Oxlo.ai: подключитесь к API, определите системный запрос, включите JSON‑вывод, добавьте функцию получения данных аккаунта, оберните логику в класс и разверните как FastAPI‑эндпоинт.

Задача

Help desks need to reduce first-response latency by automatically triaging tickets, classifying issues, fetching account data, and drafting replies.

Результат

The agent classifies tickets into Billing, Technical, Account, or General, assigns urgency, fetches mock account data, and drafts concise replies, improving response speed and consistency.

Рабочий процесс

• 1. Verify Oxlo.ai connection using a simple Llama 3.3 70B completion.
• 2. Define a system prompt that instructs the model to act as a triage agent and output JSON with category, urgency, and draft_reply.
• 3. Enable JSON mode via response_format to enforce structured output.
• 4. Create a mock tool get_account_status and allow the model to call it for account data.
• 5. Wrap the logic in a reusable SupportTriageAgent class using the Kimi K2.6 model for reliable tool loops.
• 6. Deploy the agent (e.g., as a FastAPI endpoint) and integrate with the help desk via webhook.

Доказательства

We are going to build a support ticket triage agent that reads incoming customer messages, classifies the issue, looks up account details via a mock tool, and drafts a response.

Источник

Deploy this agent as a FastAPI endpoint and wire it into your existing help desk via webhook.

Источник