Last week we focused on marketing after somewhat of a break (for exactly none of us). However, time is advancing, and so must our project in order to deliver something actually working that someone would actually use.


This week we aim to finish up a working implementation of uploading a bot to Twilio, dive deeper into React for UI, and start thinking about how exactly will we deliver something to our client.

<aside> 🎯 **What are we looking to achieve?

  1. Automate Twilio** If you've been reading these blogs carefully, you will have noticed that Twilio is giving us a hard time uploading chatbots. We then found a workaround, and now it's time to implement it properly.

2. Dive Deeper into React UI We decided to go with React (why? see below), but first, our team needs to learn how to use it properly.

3. Delivering a Product So simple, yet so difficult. We need to add clarity on exactly what will we do for our client, and until when


Automating Twilio

The way Twilio (in particular, Twilio Autopilot) works is that it allows us to upload "tasks". A task can be - ask a question, and receive a Yes or No answer.

Tasks then need to be routed - such that once one task returns a result (e.g. Yes or No), we need to start the correct next task, and so on. This is how we do conversation branching.

However, Twilio does not support routing. So we found a way to upload a custom code snippet that gets called every time a task returns, and then starts the next appropriate task. For the nerds, this is a serverless function. See how it works:

We managed to automate this whole process with a few python scripts! This was the last element of our puzzle, meaning that now we have a completely working implementation of everything, albeit not user friendly.

Diving Deeper into React UI





React is pretty much the go-to tech for doing responsive, reactive, and dynamic applications. We can easily deploy it on the web or desktop using Electron, and its component architecture allows us to remove a great deal of code duplication.