CONCURRENT SESSION 4A
Thursday 7 September 2017
Time: 3.30pm - 5.00pm
The University of Auckland
Using Machine Learning to Lower Student Contact and Support Centre Call Time
Our Student Contact and Support Centre handles a vast volume of queries, 111k calls, 216k emails and 48k face to face interactions in 2016. It takes six weeks to train a Student Advisor, of whom are often students themselves and their availability is subject to university workload. To help our strained support centre train their staff, we’ve developed a prototype bot to aid their training by bringing the knowledge to them rather than having to seek out obscure and unfamiliar knowledge base (kb) articles. We do this by allowing the Student Advisors to quickly ask the bot in natural and varied English as opposed to keywords to appease the kb search. We can then train the bot using this data to make it easier for others in the future to find the correct answer.
With this learned bot we can have it look over incoming emails for easy questions to answer such that a Student Advisor doesn’t have to spend time replying to a simple answer already presented in the student facing knowledge base as such, further reducing call (read: email) times.
I’m a software engineer from the University of Auckland working for five years for the University of Auckland under Learning and Teaching Innovation. It is my team that implements and deals with the learning management system. I do any sort of work related to things that impacts the student journey throughout the university. I’m a leadership nut and have huge lofty goals which lead me to take on all sorts of neat opportunities, such as presenting and leading exploratory projects.