Leeds Beckett University asks Alexa for help cutting Clearing admin
Dougal Scaife, head of digital experience and engagement at Leeds Beckett University, opens up about how combining chatbots with Amazon Alexa is helping update the Clearing process for the 21st century
The higher education sector is becoming increasingly commercialised, and competition between universities for students – and their all-important tuition fees – has risen accordingly, meaning institutions must do all they can to make their courses and facilities as appealing as possible to A-level students.
This is true of Leeds Beckett University in the north of England, whose student recruitment efforts focus on the use of emerging technologies, social media and digital marketing to ensure its course offerings are on the radar of prospective undergraduates and A-level students.
In recent years this work has also seen Dougal Scaife, the university’s head of digital experience and engagement, and his team move to adopt technologies that could simplify the Clearing part of the university application process for students who may have missed out on their first choice of course.
“When the A-level results come out in the middle of August, the bulk of people get accepted onto the course they applied for, but there is a whole pool of people who are left without a course, because they didn’t get high enough grades for the university they originally wanted to go to,” says Scaife.
“What happens is those individuals start ringing around and contacting other universities to see if they can get a place elsewhere.
“It is a frantic day for the students, and for us as well, as there will be thousands of people trying to find places on courses. It is in our interests to fill those places because the more people we have on a course, the more cost-effective it is to run,” he adds.
To make the process easier and less stressful for all involved, for the past couple of years the university has been looking for ways to better tailor the experience to how students prefer to access and process information in the 21st century.
“While a lot more of the [university application] processes are done online now, most universities do it the same they have always done it since the 1950s,” says Scaife.
“What we’ve tried to do is consider how people of that age, which is typically around 18 years old, use technology. We came to the conclusion that a lot of them don’t tend to use the phone that often [to make calls],” he adds.
“They would much rather communicate through digital means, using online chat-type tools, so why would we make them ring up to see if we have a place for them when alternative methods exist?”
Introducing Becky the chatbot
This realisation paved the way in 2017 for the university to trial the use of a text-based chatbot to see if it could help reduce the volume of calls to its applications team.
The chatbot, named Becky, was accessible through desktop and mobile devices, and – using a question and answer interface – was able to find out a student’s qualifications and exam results and use this to determine if they should receive a provisional offer or not.
It was also able to resolve more generic, yet pressing, queries from prospective students about elements of the application process, university term dates and their accommodation options.
“Our initial feeling was, if we programmed it well, it could bring huge benefits, as a lot of questions people ask us over the phone are straightforward and can be dealt with quite quickly, meaning our [operators] are on the phone saying the same things to people over and over again,” says Scaife.
“Our plan was to take some of the weight off the phone calls and answer the basic questions online, so the more complicated questions could still be dealt with over the phone,” he says.
By the end of the 2017 Clearing period, Becky the chatbot had notched up 21,000 conversations with prospective students, and helped 200 people secure places on Leeds Beckett University courses
By the end of the four-week Clearing period, Becky had notched up 21,000 conversations with prospective students, and helped 200 people secure places on Leeds Beckett University courses.
“Students were getting offers on courses 24 hours a day,” says Scaife. “At 3.30am, we were seeing people asking about courses and receiving offers, whereas Clearing at most universities is something that just happens when the phone lines are open, during office hours.”
The willingness of students to use the chatbot is all the more impressive because of how little fanfare went into announcing its launch. At the time, its purpose was simply to prove the concept had legs, and would be financially worthwhile to develop further.
“We developed the bulk of it ourselves, which worked out at about 100 hours of work, and we paid $30 for the platform we used, and generated £2.4m worth of potential revenue, which was a bonus for our finance director, in terms of return on investment,” he adds.
Enlisting Alexa’s help
For the 2018 Clearing window, the university has moved to expand on the initial success it has had with its text-based chatbot by extending its capabilities to include voice search and recognition by adding the Amazon Alexa virtual assistant to its technology mix.
“We had some help from the guys at Amazon, who were keen to work with us because we were doing something [in the university sector] that no one else – as far as I’m aware – had ever done,” says Scaife.
“Last year, in 2017, no one else in the world had done what we had done in terms of [using chatbots] to process applications, so we wanted to stay ahead of the curve, and part of it is to keep pushing university recruitment forward,” he says.
To get the ball rolling, the team had to build up Alexa’s knowledge of all the courses Leeds Beckett University offers, how many UCAS points are needed to get accepted on them, before training it to gently interrogate students to gather this information.
“It asks how many UCAS points a student has [based on the A-level results the student has achieved], and asks what sort of subject area they might be interested in,” says Scaife.
“It then interrogates our database, and checks to see what courses match the number of UCAS points the student has, and returns three to four courses that fulfil that criteria, and students can then ask it for more information about them.”
The technology can also indicate to students whether or not they are likely to be offered a place, before handing over to a member of the admissions team to complete the application process.
“We couldn’t go through the whole process because of the complexities of the information we had to gather, so students would get a call back within 30 minutes [of ending their engagement with Alexa] to process the rest of the application,” he says.
“To be honest, we were still building it the week before it was due to go live, so if we’d had a bit more time we might have been able to perfect that process, but our intention was to treat it as a research and development project to see if people would use that sort of thing in the first place.”
Taking the stress out of the Clearing process
At the time of Scaife’s interview with Computer Weekly, the 2018 Clearing process was still in full swing, so it will be some time before the university is able to get a full picture of how big a difference introducing Alexa to the Clearing mix has made this year, from a quantitative point of view.
Anecdotally, though, the early signs suggest the experiment has paid off on a number of fronts, as the university makes good on its goal to tailor the Clearing process more closely to students’ preferred methods of communication.
“Some people would probably look at what we’re doing as an attempt to be cool and [argue that] it’s not helping anyone, but I think this technology will be ubiquitous in a couple of years’ time”
Dougal Scaife, Leeds Beckett University
Clearing, by its very nature, is often a very fraught process for students, says Scaife, and chatbot technology has shown to help alleviate a lot of the stress involved.
“If you don’t get the results you want, it is going to be very stressful. You get a lot of students who are extremely upset. Being forced to ring up and speak to somebody can be quite an emotionally difficult thing to do,” he says, “so having the opportunity to interact with something that you don’t have to display emotion to, potentially, is of great benefit.”
So much so, says Scaife, that once the new university year is in full swing, the organisation plans to conduct a thorough review into how the use of Alexa and its chatbot could be expanded and built upon in the coming years.
“You have to keep moving forward, and we’re always looking for things we can use to ease the process as it’s important not to make technology for technology’s sake,” he says.
“Some people would probably look at what we’re doing as an attempt to be cool and [argue that] it’s not helping anyone, but I think this technology will be ubiquitous in a couple of years’ time.”
Leading the way
To this point, Scaife says Leeds Beckett University’s forays into artificial intelligence have caught the attention of other higher education institutions, keen to explore new student recruitment and outreach methods.
“There are still voices in the sector that will say [artificial intelligence] is devaluing the whole application process, [and question] how an automated ‘thing’ can make the nuanced decisions that will affect somebody’s life for the next three years,” he says.
But Leeds Beckett University’s own user data, accrued from its Becky chatbot experiment, serves to highlight just how much value the university sector could stand to gain by embracing new technologies to enhance the application process.
“What we discovered with the offers we made through the chatbot is the propensity for students to take up the offer and actually come to the university was almost double that of the people who came through the traditional phone method,” he says.
“The conversion rate for the phone method is circa 25-26% and, from the people who used the online method we saw a 46% conversion rate. It was much higher and those students turned up and are still here, and were seemingly happy with that process and that experience.
“It might not be right for everybody, but for the people who want to use it, our experience shows they are much more likely to follow through on the offer they receive and take it more seriously than the people who use the phone method,” he adds.
And, as society as a whole becomes increasingly comfortable and accustomed to using voice-activated bots to help run various aspects of their lives, Scaife says the work Leeds Beckett University is doing is also serving to ensure the institution is primed and ready to take full advantage of the trend.
“From our perspective, doing this sort of stuff now – even on a small scale – is something we can build on, year on year,” he says.
“And when it does become more ubiquitous, we will have perfected the stuff we’re already doing, and have something that is really engaging and delivering what users want, while everyone else is still rushing to keep up.”
Read more about technology in the education sector
Andrew Proctor, director of digital services at Staffordshire University, describes how ditching its datacentres and moving to the Microsoft Azure public cloud has set the organisation on the path to digital transformation.
Computer Weekly meets the Alan Turing Institute team and finds out how using cloud-first thinking can help achieve its goal of becoming a global force to be reckoned with in data science research.
