The era of Virtual Assistants and Natural Language Processing in higher education is upon on us
The number of universities and colleges developing and deploying Virtual Assistants is on the rise – there are now several UK and International institutions using this technology to allow staff and students to access core university services. Utilising a variety of technology platforms users are now able to access their timetables, view study resources, monitor attendance, book meetings and much more via simple yet powerful Virtual Assistants. The days of having to use a range of complex interfaces and systems whilst studying or working at universities are numbered.
Are there any examples of Virtual Assistants in Higher Education?
Here are some examples of the powerful uses of Virtual Assistants across the education sector.
Institution | Assistant Name | Find out more |
Lancaster University | Ask L.U. | http://itpi.lancaster.ac.uk/voice/lu.html
|
Deakin University | Deakin Genie | https://www.deakin.edu.au/life-at-deakin/why-study-at-deakin/deakin-genie
|
Leeds Beckett University | Becky | https://www.leedsbeckett.ac.uk/news/0817-becky-chatbot-clearing/
|
Bolton College | Ada | https://www.boltoncollege.ac.uk/latest-news/praise-for-ada-bolton-colleges-chatbot/
|
Stafford University | Beacon | http://www.staffs.ac.uk/news/2019/01/introducing-beacon-a-digital-friend-to-staffordshire-university-students |
How do Virtual Assistants work?
Whilst there are many technologies that underpin the development of Virtual Assistants, one of the most important is Natural Language Processing (NLP). Which is now widely available across many development platforms including Azure, AWS, Google Cloud and IBM. NLP in its most basic terms is a form of artificial intelligence that allows machines to understand spoken and written language. Integrating NLP into user interfaces has never been easier and it allows users to interact with complex systems via simple written or spoken words.
Virtual Assistants are fast becoming a method for creating a single powerful user interface that allow people to interact with many disparate backend systems. The examples above allow users to obtain information about timetabling, the VLE, library systems and attendance from one place, removing complexity and technical debt from the view of users.
Has this technology been used at UoB?
The IT Innovation and Business Partnering teams have been working closely with External Relations to pilot the use of a simple Virtual Assistant on the University website. The purpose of this prototype was to answer questions that prospective students might have before joining the university, including:
- Help with finding courses
- Help with registering for Open Days
- Information about scholarships and bursaries
The Virtual Assistant was available on several undergraduate pages on the main website for a week, with the aim of understanding the questions from prospective students. It gave us an opportunity to test the ability of a simple NLP service to answer basic questions and point users to certain resources. Some basic numbers from the pilot are stated below.
Number of Chats | Number of times it failed to answer a question | Avg. Chat Length | Pages/Session | Abandonment rate (request to speak to person) |
570 | 9.60% | 05:16 | 5.65 | 31.90% |
Whilst a further detailed report about the pilot will be published shortly, users clearly expect that Virtual Assistants should understand their context and allow them to complete more complex tasks than was possible. The pilot has been a valuable exercise in improving our understanding of NLP and how we should use the technology to design future virtual assistants.
What is the Future of Virtual Assistants at UoB?
One limitation for the development of more complex Virtual Assistants at UoB is the lack of service-oriented integration across several key university systems. Even though access and integration with NLP services has become easier, the organisations detailed earlier have stated that a significant amount of energy is required to reduce technical debt and consolidate existing services before building a highly capable Virtual Assistant.
The IT Innovation team will continue working closely with other teams across IT and Professional Services to explore the potential application of Virtual Assistants and NLP in ways that can improve the staff and student experience. There are likely to be several future opportunities to deploy Virtual Assistants at UoB powered by services such as Office 365 and ServiceNow.
How can I find out more?
If you have any questions or ideas for the applications of NLP or Virtual Assistants, please email or Teams message – Tim Packwood from IT Innovation.
A special thanks goes to the Web and Digital team for supporting this pilot.