denisismagilov - Fotolia
Lancaster University enhances student engagement with cloud-based chatbot
University is using Amazon Web Services to enhance the development of a voice-enabled “online companion” with machine learning and artificial intelligence
Lancaster University is working towards improving student engagement by developing a cloud-based chatbot using voice technology.
The university launched a trial of Ask LU, a chatbot running on the Amazon Web Services (AWS) cloud, delivered via mobile phones and Amazon smart speakers set up in the rooms of 150 students.
The project is intended to improve communication with students in general while catering for the specific needs of freshers and students with disabilities such as dyslexia or visual impairments.
Through the voice interface, students can get information about a range of subjects, from timetables, tutors and grades to how many PCs are available for use, whether a certain book is available at the library and how many washing machines are currently in use.
Chris Dixon, head of IT partnering and innovation at the university, said the project was developed in recognition that students might not want to communicate with the university in traditional ways.
“Ask LU can provide information to students who perhaps don’t want face-to-face engagement in certain situations,” Dixon told Computer Weekly at the AWS Public Sector Summit in Brussels. “Freshers might be embarrassed about asking where things are, while other students might not want to look up information on their mobiles as it can be a distraction when they are studying.”
The project went live in March following a process that lasted 120 days – from getting executive buy-in, developing early prototypes and surveying 3,000 students about what they would like a chatbot to do, all the way to developing the system, adding skills to the platform and integrating it with other functionality, such as the university’s main app, used by 98% of students.
The university will now use the AWS stack to introduce other functionalities to the platform. Dixon said the second phase of development will involve use of artificial intelligence and machine learning to get insights from data – for example, a student with poor academic performance who may not be exercising enough – and plan meaningful interventions while enhancing learning experiences.
Future functionality could also include recorded voice feedback from tutors to accompany grades, as well as more digital services and more endpoints, including Facebook. If all goes well, the voice assistant could be available in up to 6,000 student bedrooms.
“Voice technology can be a very powerful tool to deliver efficient student services and that can be done easily on the cloud,” said Dixon, adding that the idea is to make the tool feel more like a friend to students.
“Monitoring intents gave us insight around things like students asking for jokes,” he said. “We were not giving an answer to those requests, so that’s an example of things to consider for the next phase of development.
Read more about IT in universities
- How Australia’s Curtin University is transforming the digital classroom.
- University of Utah prescribes Hitachi storage for data woes.
- Australian university reaches for the stars after HPC boost.
“At the moment, the chatbot is friendly in its responses and it knows a lot of information, but it would be good to give it more depth. A friend is able to give help, advice and company and it would be nice to see this voice product doing that.
“Putting more data into Amazon or cloud-based services will be key in that process, so that we can understand complex information and distill it. Doing that, as well as having more time and resource, will enable us to make better decisions and use the chatbot to deliver better outcomes for students.”
As for key lessons learned from the initiative so far, Dixon pointed out that while 120 days to get such a project off the ground is quite a short timeframe for a university, the Ask LU project is paving the way for future innovation.
“The project has showed that even when you are constrained by processes, it is still possible to deliver something that is quite a major project in a small timescale. Technology helps, but you also need to work in an agile way,” he added.
Three full-time developers, two of them former students, took part in the project. Students from other areas, including linguistics, marketing and design, were also involved. That was another ingredient that led to a successful initial deployment, said Dixon.
“Having students involved is very useful, as they can help with quality assurance and give you lots of very straightforward feedback,” he added. “If they think something isn’t going to work, they will definitely tell you.”