2nd Place at the Facebook Messenger Hackathon at Campus Party

The team was awarded at the 10th edition of Campus Party Brazil (#CPBR10).

My name is João Dias and I was 18 at the time of the event. It took me exactly 39 hours and 59 minutes to complete a challenge launched by Facebook, the Facebook Messenger Botathon, at the 10th edition of Campus Party (#CPBR10), held in January and February 2017.

Yes, we hit the submit button in the last minute.

Diego Rangel and Douglas Lacerda (two people I met at the conference who later became friends) and I competed against 95 teams from across the country. We won second place and took home the FbStart Accelerate track code, worth US$80,000 in investment to keep developing our project: ezBOT, a bot that creates automated messages for anyone running a business page on Facebook.

Facebook had announced ahead of time that it would launch a challenge at the event, though it hadn't said exactly what. It was during Campus Party that the world's largest social network told the public what the challenge was: build a Messenger bot that solves a real-world problem and uses at least 6 of the platform's features. Teams had up to 40 hours to pull it off.

We spent nearly 40 hours at the workbenches running on coffee and energy drinks. The adrenaline hit in the final minutes; my heart clenched at the thought of not finishing in time.

The project that took second place was built around a real need of Facebook page admins who run a business. The shortest way I can put it: "it's a bot that builds bots." But to make it clearer what it actually is, take a look at how it works:

Screen showing the automated-response model, with questions and answers.

ezBOT lets a Facebook page set up custom interaction patterns with users who reach out via Messenger, like a virtual assistant. With ezBOT you can, for example, create a set of questions and answers that come up often among the customers of a given product or service. So when a user contacts the page through Messenger, they get an automated reply based on what they asked. For the user, that's a win in response speed; for the business owner too, since it makes a good impression and clears up common questions very quickly. We built two more virtual-assistant templates during the challenge as well.

Update: the project has since been discontinued.