Customer Retention System for Telecom Industry

Proactively reduce customer churn with AI-driven personalized offers

---

Description

Customer retention is a major challenge in the telecom industry, where customer churn directly impacts revenue. According to a Harvard Business Review study, retaining an existing customer is significantly more cost-effective than acquiring a new one. Inspired by this insight, this project aims to proactively prevent churn using a combination of machine learning, AI-driven personalization, and data visualization.

• Detects churn risk ahead of time with a Linear Regression model trained on the IBM Telco dataset.
• Automates personalized offers via an Agno AI agent that pulls a customer’s data, churn score and available promotions to craft and send tailored emails.
• Visualizes customer insights in a Tableau dashboard—demographics, usage patterns, churn trends—so admins spot issues and opportunities at a glance.
• Surfaces common pain points through NMF topic modeling on Rasa chatbot logs, helps to identify customer need and problems.
• Helps customers directly with a Rasa chatbot for quick answers about service issues and know about offers.

Together, these components turn raw data into proactive retention actions—keeping customers engaged before they consider leaving.

---

Table of Contents

1. Installation
2. Usage
3. Contributing
4. License
5. Contact Information
6. Acknowledgments

---

Installation

Prerequisites

• Python 3.8+
• Git
• Docker

Clone the Repository

git clone https://github.com/arjunravi26/Customer-Retention-System.git
cd Customer-Retention-System

Create & Activate Virtual Environment

python3 -m venv venv
source venv/bin/activate    # On Windows: venv\Scripts\activate

Install Python Dependencies

pip install -r requirements.txt

Rasa Setup

cd rasa_chatbot
rasa train
rasa run --enable-api

---

Usage

To run service(containers)

docker-compose up --build

---

Contributing

Contributions are welcome! To get started:

1. Fork the repo
2. Create a new branch: git checkout -b feature/YourFeature
3. Commit your changes: git commit -m 'Add some feature'
4. Push to your branch: git push origin feature/YourFeature
5. Open a Pull Request and describe your improvements.

---

License

This project is licensed under the GNU GENERAL PUBLIC LICENSE. See LICENSE for details.

---

Contact Information

• GitHub: @arjunravi26
• Linkedln: Arjun Ravi

---

Acknowledgments

• IBM for the Telco Customer Churn dataset
• Rasa for the open-source chatbot framework
• Agno AI for agent orchestration
• Tableau for data visualization tools
• Research insights from Harvard Business Review on customer retention