Link to Tutorial

https://github.com/ngrq123/bright-chatbot-tutorial

Inspiration

All of us graduated from University a couple of months ago. Between the long nights spent huddled at home from Covid-19 in the earlier stages of the pandemic to the long nights now spent on work projects as employees of various MNCs, we realised that the one thing we really missed was long nights working on fun IT projects with our mates.

What it does

Bright Social Enterprises Commerce Bot is incorporated into Facebook Messenger and utilises Wit.ai’s NLP engine to provide life-like conversations. Supplemented with Facebook Messenger’s in-built tools like shop-in chat and the ability to make payments, it provides a powerful solution to small businesses planning to build their first e-commerce chatbot.

How we built it

First, we envisioned how to best present this tutorial to an experienced learner. We wanted them to get to work on a variety of touchpoints including back-end development (node.js), building a database (MongoDB) and NLP (wit.ai). Working backwards from the learning outcomes, we came up with a list of checkpoints that the learner had to arrive at to best utilise all of the Facebook and associate tools that were employed into our solution. Ultimately, we had to find the right balance between guiding the learner and letting them explore on their own.

Challenges we ran into

  1. Virtual collaboration is a lot more challenging than working together in-person and there was increased difficulties in shared learning and collaborating efficiently.
  2. As newly-minted working adults, we could only work on the project after office hours which made it more tiring than previous competitions we had joined as students.
  3. Some of us were inexperienced with the various technologies utilised and had to take time to get familiar with what we were to use for the project.

Accomplishments that we're proud of

  1. Building the actual working chatbot!
  2. Planning a tutorial and it’s content with a new learner in mind.
  3. Crafting and working through the tutorial from the perspective of the learner.
  4. Ensuring that the content for the learner is easy to follow

What we learned

  1. How to build a Facebook Messenger Application and incorporate its tools and capabilities effectively
  2. Learned about setting up Node.js servers with webhooks to communicate with Facebook messenger
  3. Training the Chatbot using Wit.ai’s easy-to-use interface and powerful NLP capabilities

What's next for Create Smarter Messenger Experiences on Facebook with Bright

  1. Expand the used cases to larger organisations
  2. Utilise machine learning to provide personalised product recommendations based on users’ interactions on Facebook and through the chat
  3. Allow beneficiaries to insert their custom thank you note to the customers
  4. Integrate with Facebook Marketing Graph API to retrieve products from Facebook Shop to handle interactions with our application
Share this project:

Updates