Chatbot success for Erasmus University Rotterdam students

Erasmus University has several central departments that students can go to for advice and to get answers to their questions on various student matters. There is an international office for international students, a language and training center, student advisors and psychologists, and the Erasmus Student Service Center, where a team is available each day to help with practical questions. In order to deal with the increasing number of inquiries regarding tuition fees, Christa decided to look for an automated solution:

“I had heard from the Student Service Center that our employees were frequently being asked questions regarding tuition fees. The chatbot is intended for both prospective and current students. The bot is most relevant for international students, however, because the rules for tuition fees differ for this group. In addition, we wanted to gain some personal experience with using a chatbot, and to test firsthand what we could expect to achieve with automation. ”

Our young target group is “tech savvy” and can handle new technology well. Students are not afraid to use it either. Christa was once asked by a student why EUR didn’t use chatbots yet. “I think that sooner or later we will all have to deal with Artificial Intelligence. I also think that chatbots are a form of technology that everyone will ultimately use and they will become the standard”.

EUR has set a clear objective for implementing the chatbot. Measuring new forms of communication also happens to fit well with Erasmus University Rotterdam’s strategy. A minimum of 60% user-satisfaction was set to determine the bot’s success — if it had served its purpose. Christa says this KPI was more than achieved:

“About 10% of page visitors use the bot and about 80% of everyone who uses the bot also finish the conversation completely. The bot is quite complicated, it contains many questions and some questions are difficult. No less than 90% of the users are satisfied, which means that we are well above our KPI!”

There’s a marked difference in both the appreciation for and actual use of Desiderius between Dutch and international students, Christa says. The international target group uses the bot much more and at the same time appreciates it more — in both absolute numbers and in percentages. This can also be explained, because the rules regarding international students are more complex. It’s made clear by the figures presented by Christa:

“In the 11 weeks after going live, 286 Dutch students made use of the Dutch-language chatbot. The number of international students who used the English-language bot during the same period was 697. That’s almost 2 1/2 times as many. The exact satisfaction percentage, so far, for the Dutch-speaking bot is 85.6%. For the English-speaking bot it’s higher, namely 91.6%.”

Before the bot was launched, a group of EUR employees tested it extensively on both content and technology. Once the test period was completed, EUR first decided to do a soft launch. Christa: “We didn’t publicise the go-live at all. The bot was first live for three weeks before we launched it through a campaign. The funny thing was that people started using the bot right away. I’m very happy we did it that way because things came up that we didn’t see during the test period. For example, the call to action was not clear enough”.

For the PR campaign, the team looked for the best way to reach students. It turned out to be Instagram: “The bot was launched via Instagram through campus vloggers and we also posted messages on our internal channels that were read by a lot of students. There was a huge peak in the first two weeks the bot went live. After that it dropped slightly and now page visits have stabilised.”

There were certain requirements at Erasmus University that the chatbot had to meet. A number of things the team paid attention to and tested during the implementation of the bot:

• Security: the bot can of course not be hackable.
• Privacy: how do you handle requesting personal data?
• Accessibility: EUR website must be accessible for people with disabilities. The chatbot must also meet this requirement by, for example, using speech or being able to select the tab key for the visually impaired.

The chatbot was launched just a few months ago, but Christa is already thinking about other possible AI solutions within EUR. In specific areas where a lot of inquiries made, it seems quite logical to use bots:

“What we’re thinking about now is how we can continue to develop this bot. We’re investigating whether we can optimise the chatbot even further and perhaps even include more AI in it. An example of this would be where users would not only make a choice via buttons, but also by entering text in a text field. What would also make sense would be to use a bot at the back and front end of the service center. In the back-end, for example by labeling questions, and in the front-end by providing standard answers to simple questions. I can also imagine that at some point we will think further about implementing a choice of study bot.”

As advice for other universities or schools, Christa recommends starting with a well-defined topic: “There are so many fun things to think of that you can tackle with a chatbot. But if you try to tackle it all immediately and in a big way, it becomes complicated. Start small and keep things clear and organised. First gain the necessary experience and then continue to develop and expand from there.”

Christa has experienced firsthand that there is still a lot of confusion among colleagues about what a chatbot can do exactly, and that, internally, there are some unrealistic expectations. “I noticed that when I started working with the chatbot, some people were a bit hesitant. What does this mean and what exactly will the bot do? In the beginning I still had to do some mission work in order to make the bot better known and more familiar within the university. I did that by doing a lot of talking and explaining, and showing examples. Spotler Engage was extremely helpful with this. There were steps that I myself would have easily skipped, but they were crucial because it’s still a relatively unknown technology.”