CustomerCareBot is a Streamlit web application designed to automate customer service responses using AI-powered language models. It uses your knowledge base i.e (a companies data) to generate a response. In this implementation it uses past data between customers and customer care agents. It provides a convenient interface for generating automated responses based on customer inquiries or messages.
The data was extracted from Bitext - Customer Service Tagged Training Dataset for LLM-based Virtual Assistants
-
Automated Response Generation: Utilizes OpenAI's GPT-3.5-turbo model to generate context-aware responses to customer messages.
-
Conversation History: Maintains a history of user interactions and bot responses for reference and continuity.
-
Easy-to-Use Interface: Simple text area for entering customer messages and a button to generate responses quickly.
-
Customizable Settings: Options to adjust response generation parameters like temperature and model selection.
To run CustomerCareBot locally, follow these steps:
-
Clone this repository:
git clone https://github.com/Tobsky/CustomerCare_Bot.git cd your-repository
-
Install dependencies:
pip install -r requirements.txt
-
Set up environment variables:
Ensure you have set the necessary environment variables, such as OPENAI_API_KEY for OpenAI authentication.
-
Run the Streamlit app:
streamlit run app.py
-
Open your web browser and go to http://localhost:8501 to view the app.
-
Enter a customer message in the text area labeled "Customer message".
-
Press Ctrl+Enter button on your keyboard to trigger the AI model for generating a response.
-
The generated response will be displayed in the interface, along with the previous conversation history.
-
Use the conversation history to maintain context and continuity in customer interactions.
Here's how CustomerCareBot can be used:
-
Customer Input:
Hello, I need help canceling an order I made.
-
Generated Response:
Assistant: I understand your concern. To cancel your order, please provide your order number and contact information so we can assist you further.
Contributions are welcome! If you have suggestions, feature requests, or bug reports, please open an issue or submit a pull request.
This project is licensed under the MIT License.