Dude, Where’s My Car?

AI-Powered SMS ChatBot

Description

This project is designed to be an AI-Powered ChatBot that can interpret a customers SMS message and respond accordingly. Its purpose is to ease the workload of customer service representatives from having to repeatedly answer basic questions. Examples of it working are available below
The AI utilizes Amazon's Lex, which is uses "natural language understanding (NLU) to recognize the intent of the text." Lex then interacts with AWS Lambda which decides what to do with the users intent. If the intent isn't understood, it will ask for clarification. Lex will then respond by either asking for clarification or respond with the information the user asked for.

Inspiration

The Autonation path required a chatbot to be developed. We initially considered creating a neural network that would learn each possible question that the user could ask, but we ended up going with Amazon Lex and using Amazon Lambda to handle the logic for Lex.

Challenges

  • Learning Figma
  • Learning Amazon Lex
  • Utilizing Lex V2 with blueprint lambda code
  • Difficulty teaching the chatbot specific questions

Looking Forward

The future of this project will include increasing the number of questions that the chatbot can respond to and answer. The focus will be on integrating lambda with JSON files or a database to respond to questions more dynamically. The goal would be to put the bot into an app for direct user interaction and have it work with multiple forms of communication. Such as SMS, email, chat, and voice calls.

Demo

The video demo can be found here and also on the Github repo. https://youtu.be/F-xsTgJyZ-U

Example

Prototype of full developed bot

A gif of the prototype in action can be found in the Github repo.

Current Product

The Images are above and can also be found in the Github repo.

Created By

Jairo Garciga - LinkedIn | Github
Rachel Mathews - LinkedIn | Github
Ryan Smith - LinkedIn

Built With

Built With

Share this project:

Updates