Because of COVID-19, everyone needs to minimize the amount of social contact and refrain from going to stores when possible. However, there are certain necessities such as food and water, where people have to risk infection and go to go to the store anyways. To help with social distancing, there are other options to get groceries and/or order food to go online, but for many users these sites are not intuitive and difficult to navigate. We wanted to create a more intuitive approach to get food by using chatbots to simplify the process of ordering food online. The bot would as sort of assistant with ordering food where it is more of a conversation, then trying to navigate a website on your own.

We wanted to try using the Twilio API because they had some interesting communication tools for messaging and AI bots that we wanted to explore.

What it does

To simplify the process of ordering food online, we attempted to create a bot to assist the user in completing an order. The user would type a message to the bot on the web app, and the bot would respond with the appropriate questions or comments to help the user complete their order for food.

Unfortunately, we were unable to make it fully functional. We created a chat platform using the Twilio Chat API, where the user can type in a message and it would appear in a text bubble. We also used the Twilio Autopilot API to create a bot that takes a food order, but we were unable to combine these two things into one web app where the user can actually submit a food order.

How we built it

We used the Twilio Chat API and Twilio Autopilot API to try to create a chatbot. With the Twilio Chat API we were able to create a chat platform where the user can type and send their response in the chat. With the Twilio Autopilot API, we created a bot that would respond to the user when the user types in certain questions or comments.

Challenges we ran into

We were unable to figure out how to make multiple channels for messaging. As a result, we have a strand of all the messages we typed to test the code, so when the user logs in, they are not able to only see their messages. Instead, they can see every message we typed as well.

We were unable to take the bot we made and add it to the chat platform.

We tried to make the text bubbles look more like speech bubbles by adding a triangle like shape to the end, but we were unable to get the triangle positioned for the user's speech bubbles.

Accomplishments that we're proud of

No one on the team had ever used either of the Twilio API's we attempted to use, so we're proud that we tried something new and learned from this experience.

What we learned

We learned generally how to make a chatbot. Even though we weren't successful in finishing the project, we learned what things we need to consider and generally what steps we need to take.

We learned more about the Twilio APIs, so in case we want to create something involving communication for another hackathon, we are more familiar with some of the tools out there.

What's next for chumbot

We would want to combine the work we did with the Twilio Chat API and Autopilot API, so it is actually a functioning chat bot.

We would also want to allow the user to submit actual food orders for grocery stores or restaurant take out orders through the chatbot.

Built With

Share this project: