Inspiration
With the rise of conversational AI, we can extend this technology to our neighbors in New Jersey who have trouble navigating websites. We personally use NJ Transit to travel, and although the website has great features, we believe our chatbot would be a welcome feature to the user interface.
What it does
Answers questions about ticketing, routes, bus and train stops, and account information.
How we built it
Trained a conversational AI model using open-source google cloud Dialogflow, to take in user text input based on various NJ transit data and previous responses from past users. We then integrated this model in the form of a chatbot to our frontend interface using Express.js.
Challenges we ran into
Integrating the conversation AI into the frontend interface. The lack of time in training the model was a drawback.
Accomplishments that we're proud of
Being able to ask specific questions such as refunding, ticketing, and discounted prices. Also, the chatbot's ability to handle human error.
What we learned
How to integrate advanced API into frontend frameworks. Also, training a conversational AI was very fulfilling.
What's next for Nate the Transit mate
Being able to actually book tickets, using NJ transit's preexisting database. Using customer information/location to figure out routes,prices and more.
Log in or sign up for Devpost to join the conversation.