by George Nott

University of Adelaide employs chatbot to handle ATAR score queries

May 02, 20185 mins
Artificial IntelligenceBusiness IntelligenceCloud Computing

The University of Adelaide has piloted a chatbot to handle the tsunami of enquiries that come at the end of the academic year from prospective students regarding ATAR bonus points.

The Oracle, cloud-based bot – implemented by Rubicon Red – was trialled late last year and led to significant reduction in telephone calls to the university’s six-person admissions team.

A student’s ATAR(Australian Tertiary Admissions Rank) is a score between zero and 99.95 that tells you where you rank in your year group.It is based on their HSC results with different weighting given to various subjects, and the figure is a percentile position relative to all the other students in their year.

Universities set a minimum ATAR value for each course. Students may be eligible for adjustment (or bonus) ATAR points.

There could be adjustment points added to the ATAR scores of students that do well in HSC subjects relevant to their desired university course; if their school is in the vicinity of the university or a regional area; or if they have been affected by disrupted schooling, financial hardship, illness, disability, refugee status or socio-economic disadvantage.

“During such a stressful and anxious time, we don’t want applicants having to wait in long queues on the phone to speak to us. It’s important that they can get their adjusted ATAR score as quickly as possible and start discussing what their results mean for them with their families, their friends and university of choice,” said University ofAdelaide’s associate director, prospect management, Catherine Cherry.

The chatbot – which was built within four weeks – resides within Facebook Messenger, and was trained on questions used in previous years to help determine a student’s adjusted score. Given deep learning-based natural language understanding (NLU), the bot is able to provide ‘natural, conversational and personalised responses’ including congratulatory comments.

Oracle said that machine learning helped the bot to ‘pick things up and learn as it goes along’, making it more powerful as usage increases.It also remembers previous conversations and their context.

“Our Adjusted ATAR chatbot meant students didn’t have to wait – in the busiest hour we had approximately one user every five seconds. The response has been fantastic, with more than 60 per cent of users rating their interaction as ‘Awesome’,” Cherry added.

On its first day in action the bot handled an estimated 2,100 unique conversations with the chatbot, a number far greater than anticipated.This led to a 40 per cent reduction in calls to the University on results day – and a 47 per cent drop in calls during the first three hours after ATAR scores are published.

Telephone hold times fell from an average of 40 minutes down to about 90 seconds.

“Previously, the University’s incoming student call volumes with ATAR queries were so high, the customer call centre couldn’t field them all.Now they can and have time to turn their attention to other areas. After all, the University has a team of around six people nurturing around 70,000 students. It was a really satisfying experience to help it try something different and get such a great result,” said Franco Ucci, senior director, Oracle.

The university – which appointed former SA Water CIO Bev McQuade as its tech chief in February last year – is now eyeing other areas where chatbots could be applied including around pre-application engagement with international students.

Other universities are also reaping the benefits of chatbot technology. In February OpenUniversitiesAustralia (OUA) partnered with LivePerson to roll out bots to handle processes such as username and password resets, and to provide after-hours assistance to current and prospective students, via SMS and Facebook Messenger.

“Messaging aligns us with how our students like to communicate in today’s digital world,” said Lyndon Summers, OUA’s operations manager, sales pipeline.

“They can get the assistance they need from our offices on their own schedule and terms. And they seem to prefer it over voice calls– we’ve seen a high customer satisfaction score with bot conversations of over 80 per cent on average since implementation and an ROI of over 220 per cent within six weeks of launch.”

Deakin University has its ‘Genie’ chatbot to provide personalised student advice, which will be rolled out in full in July. Late last year the University of Canberra unveiled two Microsoft Azure based chatbots to handle IT requests from students and staff. Meanwhile, the University of Sydney in March revealed it was piloting chatbots to remove the administrative burden on its finance department by handling queries regarding purchase orders, invoices and supplier-relations.