chandraashekar / hackuta

Inspiration

Our primary inspiration was to create an intelligent system that would enhance customer relationships and reduce wait times.

What it does

It bridges the gap between agents and customers using Artificial Intelligence.

How we built it

Firstly, we designed tall the possible queries that a user could come up with, that could be answered by a chatbot. We started to train our NLP model with all the possible utterances, and intents. We used LUIS.AI an Azure ML service to train our model. Then later on we created a webserver with REST API and Webhooks using node.js and express.js framework. We have used TWILIO to address user queries in message format(SMS, WhatsApp).

Challenges we ran into

Creating a mock database for training and testing purposes was a challenge as we had to come up with all possible permutations that a user could come up with. Integrating TWILIO and SMS service with our chatbot.

Accomplishments that we're proud of

Within a span of 1 hour, we have extended our service to SMS; which initially was limited to WhatsApp. Creating an NLP that understands just like a human.

What we learned

We have learned time management and to push our limits by staying up late at night and being able to complete the project in the stipulated time.

What's next for State Farm Chatbot

Platforms could be extended even more to reach a larger number of customers. Implement Image Processing/Audio clip analysis for case analysis and instant resolution.

Built With

• azure
• express.js
• git
• json
• luis
• ml
• mongodb
• node.js
• rest-api
• twilio

Try it out

GitHub Repo

hackuta's Issues

Intents