Jim Cook, Christine Tsintarakis
Martin Ashton Jones
25 August 2017
Completed 18th October 2017
What is the simplest and most cost effective way to implement Cognitive services to reduce impact and load on First level support teams at the University
A fully deployed and functional QnA natural language chat bot, deployed in Microsoft Teams, Telegram. and Skype. It answers the 150 most common First level support questions based on data from trackit and participation by Field Services and Helpdesk total build time was less than 7 days.
C1. The longest part of this project was deciphering the data. Ask Sydney is currently a black hole of walls of text.
L1. More structured approach to what is in scope and out of scope. A pilot could have been deployed using the poor data that existed, however it was decided early on that it was better to spend some of our time triaging the data so that the end product could demonstrate its value more powerfully.
C2. Getting First level support staff to work on this project took a longer time than we hopped and pushed the start date back so that it overlapped with team annual leave.
L2. Luckily the team on this project were excellent at independant work and decision making and delivered the content regardless of the time line.
Node.js, Azure, Bot Framework, Cognitive Services
Bot framework source code
Please see a Techlab team member for a Teams link to test the bot!