WhatApp HR Bot (Twilio AI Challenge)

This is a Python application that allows you to load a CSV file for employee FAQs and ask questions about it using natural language. The application uses a LLM to generate a response about your CSV File and then sends to Whatapp via the Twilio API

How it works

To build an HR WhatsApp chatbot using Twilio API and a custom knowledgebase in Python with LangChain, you need to follow these steps: Set up Twilio for WhatsApp: Create a Twilio account and set up a WhatsApp Sandbox. Get your Twilio Account SID, Auth Token, and WhatsApp-enabled phone number.

Build the chatbot:

• Set up the Flask web framework to handle incoming messages.
• Use LangChain to process incoming messages and generate responses based on your custom knowledgebase.
• Use the Twilio API to send responses back to the user via WhatsApp.

Installation

To install the repository, please clone this repository and install the requirements:

pip install -r requirements.txt

You will also need to add your OpenAI API key to the .env file.

Usage

streamlit run app.py

Deploying the chatbot

• Run the Flask app locally:

python app.py

• Use a tool like ngrok to expose the Flask app to the internet:

ngrok http 5000

• Configure the Twilio Webhook: In the Twilio Console, navigate to your WhatsApp Sandbox settings. Set the "WHEN A MESSAGE COMES IN" URL to your ngrok URL, e.g., http://.ngrok.io/whatsapp. Now, when you send a message to your WhatsApp Sandbox number, the Flask app will handle the incoming message, process it using LangChain, and respond accordingly.