At work, we love when our coworkers bring us lunch. Yay! But when we get lunch for them, do we get the appreciation we deserve? It's a big pain in the butt to ask a bunch of people what they want, listen to them debate, and then parse their answers into a simple order list. And whose turn is it to get coffee anyway?
What it does
LunchBot solves these pains by handling order requests in natural language conversation right in the team conversation on Glip. LunchBot announces that a trip is being made, takes a limited number of orders, and then sends a single text message to the driver who's already on his way to In-N-Out (or wherever!). When the driver gets to the counter to order, everything is in one simple list. Wow, that's great!
How we built it
We set up a Node server that hosts the chatbot logic built with the Botkit framework. LunchBot listens to the Glip channel and responds accordingly. We used a variety of tools such as postman, requestb.in, ngrok, and others to develop.
Challenges we ran into
Using HTML5 media capture capabilities proved to be a challenge when integrating directly with Voicebase. We found a solution to go to an intermediary server before hitting the API.
Accomplishments that we're proud of
In this hackathon, Ray learned a lot about working with APIs, having primarily a UX and UI background. Sasi learned about Nodejs and now we have a lot more confidence with full stack prototyping.
What's next for LunchBot
Imagine you'd like to interview for a promotion at work. The hiring manager doesn't know you super well, but she can see that you get your co-workers lunch more than anyone else in the company -- she even remembers that you brought her lunch just last week. Wow!