A fun side project I participated in for SleepScore Labs Summer 2019 fitness program – a 5 week total body workout incorporating cardio, calisthenics while integrating better sleep habits into your overall health and wellness routine. Are you up for the challenge? Sign up now!
A video walkthrough of an Angular Data Visualization Web App I developed for SleepScore’s App & Max Sleep Improvement Systems. The end user is able to look at additional data sets pertinent to their previously recorded sleep sessions.
The original write-up was written in Angular 2. Feel free to review it here.
In lieu of SleepScore Labs newest sleep solution, SleepScore App, our Customer Service team wanted to create a chat alternative to support this product launch.
Over the course of three furious weeks (one to vet out the needs, and two to build), I worked one-on-one with our Customer Service Manager on outlining the bare minimum of what we’d really want in this type of feature.
It’s safe to say, that you can get trapped inside of a rabbit hole fairly quickly with ideas but after easing in a bit our solution was to design an FAQ (Frequently Asked Questions) Chatbot.
We took as much existing data we could from previous customer service engagements, organized them a bit and indexed them within a Web Service called QnA Maker. I connected that dot to a Microsoft Azure Bot Service as a NodeJS Web App and published it to a channel called Live Assist.
A NodeJS chatbot powered by QnA Maker, Microsoft Azure and channel, Live Assist.
Housed entirely inside of Microsoft products, I was for the most part impressed on how straightforward it was to put all three together to build out something real customers could actually engage with. There was no additional coding on my part and getting to the finish line on time was in itself a win.
If you ever wanted to try building this out yourself here’s the bare minimum you’d need:
The optional requirement here is the channel. Building out a chatbot requires you to point your chatbot to a channel. In our case it was Live Assist because that was the requirement. But this could easily ship to other channels like Facebook, SLACK, etc.
I’ll be making a follow-up post to this for the most part as a retrospective to outline the journey and its gaps to this implementation in hopes that one day I have the opportunity to make more improvements at this first pass on building out a chatbot.
Thanks to, Microsoft Developer US for getting me started.