There’s no denying that chatbots are changing everything from how customers contact support to the way companies manage sales pipelines. And now, thanks to a partnership between Microsoft Australia and Antares Solutions, chatbots are also shaking up the local education sector. Here’s a sneak peek into an exciting new chatbot pilot project underway at the University of New South Wales.
Imagine a chatbot that knows when your next university assignment is due, which classroom your tutorials are in, and the weighting of your final exam. Now imagine that unlike regular chatbots, you don’t need to address the bot directly to have these questions answered. Instead, the chatbot proactively listens to conversations within your course collaboration platform, and provides answers when it identifies a need.
For students at UNSW’s School of Mechanical and Manufacturing Engineering, there will soon be no need to imagine. In fact, this is the ultimate vision for a chatbot that we’re working with Microsoft Australia to develop as part of a larger Microsoft Teams pilot project.
How the project came about
UNSW is already working with Microsoft Australia to trial Microsoft Teams. If you haven’t heard of the platform, Teams functions like a combination of Facebook and Slack, but for organisations. People are brought together through a simplified communication and administration system that combines chat, meetings, content, notes, attachments and more, all in one location. UNSW is already seeing positive results, with the exam pass rate for the class piloting Teams increasing from 65 percent to 85 percent.
But UNSW has bigger plans for Teams – and that’s where Antares comes in. In late 2017, Microsoft Australia recommended UNSW engage Antares to develop a Microsoft Teams chatbot that would add even more value. Antares has a reputation as a trusted leader, both in technology solutions for the education sector, and in building chatbots for non-typical applications.
Phase one of the chatbot project will soon be put into action as a pilot program to UNSW’s School of Mechanical and Manufacturing Engineering. The goal is for this chatbot to proactively listen to all student conversations taking place within the course’s Microsoft Teams platform. It will gather and identify questions from students and pass the questions on to lecturers, tutors, or even other students, to answer.
Phase two will push the bot even further. UNSW’s Dr David Kellerman, who is managing the Microsoft Teams pilot project, says his vision for the chatbot is simple.
“I want the bot to be part of the general channel, rather than being exiled to a chat. The bot needs to listen to every post in the general channel, attempt to parse the post, and if it identifies a question then attempt to answer it,” he said.
This differs from your average bot that only replies when asked a question by users.
At first the chatbot will be trained to respond to simple questions, and then as it progresses it will have the ability to answer more complex and technical questions. It almost goes without saying that a bot with the ability to identify, understand and respond to questions in the same way a tutor or lecturer does is revolutionary in the education sector.
We expect that the new chatbot will make the lives of both UNSW students and staff easier. The bot will be a huge time saver, supporting instant communication 24 hours a day, seven days a week. It will also reduce the number of emails between students and staff, cutting down on inbox clutter and the risk of messages being missed among hundreds of others.
Instead of students wasting valuable study time waiting for email responses from tutors or lecturers, the bot will eventually provide instant answers. (In phase one, questions will be answered by knowledgeable staff and other students, which is still faster than going back and forth via email.). Staff will save time previously spent replying to emails, often answering repetitive questions. The system also supports a self-learning community by allowing fellow students to answer each other’s questions.
Importantly, all questions gathered by the bot are stored together. Tutors and lecturers can note which questions are frequently asked and identify common areas where students may be experiencing difficulties. The bot’s simple feedback system, whereby users can select whether its response was ‘useful’ or ‘not useful’, also helps the bot to improve its language model and knowledge base over time.
This project is advancing well and it’s safe to say that UNSW is excited about our involvement. We are proud to assist UNSW with its vision for an intelligent bot by providing clear direction, a tangible solution and deep technical expertise. We look forward to helping UNSW to build a proactive and self-learning bot that will improve administration, productivity and communication within the university.