UtkarshTheAIMaster/openenv-ticket-routing-ai
0
title: OpenEnv Ticket Routing AI emoji: ๐ colorFrom: blue colorTo: indigo sdk: docker app_port: 7860 pinned: false
๐ Ticket Routing Environment (OpenEnv)
๐ 1. Overview
This project implements a real-world AI environment where an agent performs customer support ticket triaging and resolution.
The agent must:
- Understand user complaints
- Classify tickets into categories
- Assign priority levels
- Generate a response
- Decide whether escalation is required
This environment follows the OpenEnv specification and is designed for training and evaluating intelligent agents (LLMs/RL agents).
๐ฏ 2. Motivation
Customer support automation is a real-world high-impact problem. This environment simulates realistic workflows used in:
- SaaS companies
- E-commerce platforms
- Tech support systems
It helps evaluate how well an AI agent can make multi-step decisions under constraints.
๐ง 3. Environment Design
๐น Action Space
๐น Observation Space
๐น Reward Function
- Range: 0.0 โ 1.0
- Partial rewards for:
- Correct category
- Correct priority
- Correct escalation decision
- Meaningful response
- Penalizes incorrect or poor actions
๐งช 4. Tasks (Difficulty Progression)
โ Tasks increase in complexity and require better reasoning.
โ๏ธ 5. API Endpoints
๐ฅ๏ธ 6. How to Run (Local Setup)
Step 1: Clone Repository
git clone <your-repo-link>
cd ticket-routing-envStep 2: Install Dependencies
pip install fastapi uvicorn pydantic requestsStep 3: Start Server
uvicorn my_env.server.my_env_environment:app --reloadStep 4: Open in Browser
http://127.0.0.1:8000/docs๐ Use Swagger UI to test endpoints interactively.
