Amazon Hackon 2023 (Season 3) Prototype

Theme: Contextual Shopping experience with Generative AI + AWS

Team Name: Techno Sapiens

Team Members: Aryan Singh (Team Lead), Chanpreet Singh, Himanshu Upreti, Nikesh Kumar

Prototype Resources

• Deployed on Website: Website Link
• PPT of our prototype: Canva Link

Prototype Video

• Video of our prototype: Youtube Link

  

Prototype Images

1. Contextual Shopping Chat Bot

   

   

2. Chat Bot storing contextual preferences and giving results based on previous conversations

   

3. Post Purchase Chat Service Bot

   

Introduction

Welcome to the Amazon Customer Assistant Chatbot project, your friendly virtual shopping companion! In online shopping, we understand that finding the right product can sometimes be overwhelming. Imagine you're in a physical store looking for a TV, and a knowledgeable salesperson is there to assist you. They consider your room size, budget, and preferences to guide you towards the perfect choice. This personalized approach is what we call "Customer Obsession." However, in the online shopping landscape, it often feels like you're navigating the vast digital aisles all by yourself. That's where the Amazon Customer Assistant Chatbot comes to the rescue. Our mission is to replicate the experience of having that friendly salesperson but in the digital realm. We are genuinely obsessed with ensuring your satisfaction as a customer.

Project Overview

Our Goal Our primary objective is to provide you with expert advice, recommendations, and suggestions, so you can confidently make the best choices when shopping on Amazon. Whether you're looking for a TV or any other product, our chatbot is here to simplify your online shopping journey.

How It Works

Our chatbot is powered by a state-of-the-art Language Model (LLM) based on OpenAI API. When you interact with our chatbot, it uses this advanced model to understand your queries, preferences, and needs and generates a query based on all that information then we do a vector similarity search with our database based on the query generated by our LLM model. It then provides you with personalized, informative, and relevant responses, by matching just like a seasoned salesperson would in a physical store.

Features

1. Personalized Recommendations: Our chatbot considers your individual requirements, such as budget, room size, and features, to suggest products that best match your needs.
2. Real-Time Assistance: Please get immediate responses to your questions, making sure you have the information you need when you need it.
3. Expert Guidance: Our chatbot leverages the collective knowledge and data available on Amazon to provide you with expert insights and advice.
4. Easy and Fun Shopping: We aim to make your online shopping experience enjoyable, just like having a knowledgeable friend guide you through a store.
5. Provide personalized product recommendations based on customer preferences.
6. Implement a similarity search using vector embeddings for a better shopping experience.
7. Utilize the MERN (MongoDB, Express, React, Node.js) stack for development.

Why Choose Amazon Customer Assistant Chatbot

1. Customer-Centric: We are dedicated to your satisfaction, and our chatbot is here to assist and empower you in your online shopping journey.
2. Efficiency: Save time and make informed choices with our fast and accurate responses.
3. Expertise: Benefit from Amazon's collective expertise and advanced AI technology capabilities.
4. User-Friendly: Our chatbot is designed to be user-friendly and accessible, so you can shop with ease.

Technologies Used

• Frontend: React
• Backend: Node.js, Express.js
• Database: MongoDB
• Vector Embeddings: Used for similarity search
• LLM Model: OpenAI API
• Other Technologies: Selenium (for automation and web-scrapping), AWS EC2 Instance (for hosting)

Methodology

Installation.

1. Install the MongoDB Community server. Link
2. Download and Install Visual Studio 2022. Make sure to select "Desktop development with C++". Vs Code Download link
3. Clone the repository git clone https://github.com/chanpreet3000/amazon-hackon
4. Create a .env file in ./backend and paste the following text

     REACT_APP_BASE_URL = "http://localhost:8000"
   

5. Create a .env file in ./frontend and paste the following and set the details accordingly.

     JWT_KEY = "TEMP_JWT_KEY" # you can cahnge accordingly
     MONGO_URL = 
     OPENAI_API_KEY = 
     PORT = 8000
   

6. Open the repository using a code editor and run the following commands.

     cd ./backend
     npm  i
     npm run start # start the backend
     cd ../
     cd ./frontend
     npm i --legacy-peer-deps
     npm run start #start the frontend