customer-support-agent

A production-ready customer support agent using Weave and W&B Inference.

Overview

This project implements a customer support agent that leverages Weave and OpenAI's capabilities to provide efficient and safe customer interactions. It includes guardrails to ensure account and content safety, making it suitable for production use.

Tutorial Reference

This repository accompanies a tutorial article. You can read it here.

Prerequisites

• Python 3.8 or higher
• Weave library
• OpenAI API key (create an account at OpenAI)
• W&B account (create an account at Weights & Biases)

Quick Start

1. Clone the repository:

   git clone https://github.com/yourusername/customer-support-agent.git
   cd customer-support-agent

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Create a .env file in the root directory and add your API keys:

   echo "WANDB_API_KEY=your-api-key" >> .env

4. Run the application:

   python src/main.py

Project Structure

customer-support-agent/
├── .gitignore
├── requirements.txt
└── src/
    ├── agent.py
    ├── guardrails.py
    └── main.py

Key Concepts

• Weave: A framework for building and deploying machine learning applications.
• Guardrails: Safety checks implemented to ensure that the agent operates within defined safety parameters.

Code Highlights

• Guardrail Implementation: The FlaggedAccountGuardrail class in guardrails.py demonstrates how to implement safety checks for flagged accounts, ensuring that sensitive operations are restricted.
• Asynchronous Operations: The use of async in run_guarded_agent in agent.py allows for non-blocking execution, improving the responsiveness of the customer support agent.

License

This project is licensed under the MIT License. See the LICENSE file for details.