Towards the end of the 3rd quarter of 2021, the Leadership Team started talking about how to handle better several of the IT issues we had. On one hand, we aim to have a production that is undisrupted by any system issue. On the other hand, we always want to keep away from the dreaded IT issue escalation from any client.
At that point, the initial IT issues escalation matrix was drafted and it went thru several discussions on how to improve it. One of the initial ideas was having a chat station or a chat desk that would be interactive in providing the info on managing IT issues. We initially thought it is a simple FAQ project where we simply build a table that maps corresponding responses to every IT issue. But as it turned out, we discovered it involved machine learning.
Machine learning or ML is not something new. It has been there since the 1950's when an IBM programmer built a checker game that got better as it faced more opponents. However it's has been only used widely by major companies starting 2010. Somehow I felt it's still something unexplored by most. And you may not know it but you may already be using it every time you say, “Alexa, play Spotify.” or “Siri, tell me a joke.” When you watch Netflix, you would be amazed at the recommendations that it's already giving you. That is machine learning.
Back to the project, the Imagineer team started to spend time to learn about how we can set-up the machine learning with the chat application we’re using, Webex. After about two weeks of back-and-forth discussion and video calls with the Google Cloud team, we were able to establish the connection between Webex and the Google platform we’re using, Dialogflow and complete the set-up.
The next phase was the training phase. The first time the team introduced the Cooee chatbot to the Leadership Team, it only knew how to respond properly when you said, “Hi” or “Hello”. After that, it does a loop of asking you, “Say that again?” or “One more time?”. As of today, the Cooee chatbot knows about 50+ intents or categories of conversation, about 150 scenarios of a conversation and about 700plus training phrases.
Why do we do what we do?
From the business side of things, here are some points.
Economies of scale: As it has been so since start of last year, Cooee Inc. is expanding. The company is growing. We want to ensure that the level of IT support or coverage is still there 100%, despite the small IT team we have for now.
With this chatbot, basic queries are addressed and we ensure that there is no bottleneck with those queries. Everyone can ask the chatbot at the same time and it will still be able answer everyone at once.
This ML tool captures data and it captures all queries so that in the long run, this tool would be able to give us insights on the kind of issues we are experiencing therefore bringing us to better directions, better business decisions.
From the employee or user’s perspective, here are some benefits.
Cooee iTee can be your personal assistant. It is the little far cousin of Alexa or Siri. If you're searching for answers for your whole life, i.e. IT related ones, the Cooee iTee is there to give you answers.
The tool is available 24/7.
Solutions or answers are customized according to Cooee’s needs and values, because this chatbot is trained by Cooee, by everyone who took time to drop the bot several test messages during the testing phase.
Comments