E-Commerce Chatbot — AI Assistant for Policies & Products

Live App: https://e-commercechatbot.streamlit.app/

This is a generative AI chatbot designed for e-commerce platforms. It can answer both general FAQs (e.g. "What is your refund policy?") and product-specific queries (e.g. "Show me PUMA shoes above 30% discount") using a combination of vector search, LLM-based reasoning, and SQL generation.

It uses Groq's LLama3.3, ChromaDB, and Streamlit to deliver fast and accurate responses.

---

How It Works

1. User enters a query (e.g., “Do you accept debit cards?” or “Top 3 Nike shoes under 3000”)
2. The query is passed through a semantic router, which classifies it as either:
   • faq → answers are retrieved from a vector DB (ChromaDB)
   • sql → the question is translated into an SQL query and run on the product database
3. The appropriate chain (faq_chain or sql_chain) is called.
4. The result is passed to a Groq-hosted LLM, which generates a conversational response.

---

Demo Screenshots

Chatbot in Action

Architecture Flow

---

Getting Started

1. Install dependencies

pip install -r requirements.txt

2. Add your .env in the app folder

GROQ_API_KEY=your_actual_groq_api_key
GROQ_MODEL=llama-3.3-70b-versatile

3. Launch the app

streamlit run app/main.py

---

Notes

• The product data comes from a one-time web scraping of Flipkart using Selenium.
• The app does not generate responses unless a valid GROQ API key is provided.

---

Contact

Kushal Samani

• Email: kushalsamani04@gmail.com
• LinkedIn: https://www.linkedin.com/in/kushalsamani