NCOI introduces chatbot for optimum service for students

The team sought a solution for answering frequently asked questions in the student portal. An important objective was to be able to talk to the students outside of opening hours. In this interview, you’ll read all about how, despite internal challenges, the team successfully launched the chatbot “Joost”!

The bar has been set high for chatbot Joost

Chatbot Joost can be found in the learning environment of students following a course taught by one of the NCOI Group’s trainers.

Chatbot knowledge-owner, Max Defize, has been closely involved in creating conversations for the bot:

“You see that difficult questions, such as re-taking an exam, can easily be answered by a chatbot. It’s nice to see that, during a conversation, the chatbot can sometimes find solutions all by itself. On busy days, the bot can be clicked on around 500 times.”

The complete team of study coaches at NCOI consists of around 80 employees spread out over several teams. From customer advisors to counsellors who tackle specific cases. For the most part, the transition from chatbot conversations to live chat conversations goes quite smoothly. As a ‘work planner’, the chatbot does a lot of the preliminary work, such as answering certain questions, before the conversation is transferred. “We see a simple question on the back end come through, such as: ‘When will I get my study materials?’ and that saves a lot of phone calls to students. These are great things to see”, explains Knowledge and Quality consultant Lotte Vermeulen.

Qualitative conversations despite employee turnover

The main reason for starting with a chatbot was to be able to offer good service. A sub-goal for NCOI is that the answers are always good ones, and qualitative in terms of content. Eric explains: “Like every customer contact centre, we have to deal with turnover, so information comes and goes. We strive for quality customer contact, and guaranteeing substantive knowledge has a positive effect on this. The chatbot helps to maintain that quality”.

In addition to the large turnover among employees, the complexity and size of NCOI is also a challenge. Lotte explains: “We have 13 different trainers within the NCOI Group that we work for and therefore so many different target groups, regulations and agreements. And then you have all the topics that you want to add to the chatbot. From exams to study materials to teachers. That makes it quite extensive”.

Based on the complete chat history data set, Spotler Engage did a feasibility analysis for NCOI. For this purpose, in addition to existing conversations, all the information in a knowledge base of the most frequently asked questions was also used.

Implementation and going-live

Colleagues from all around the organisation were involved in the implementation of the chatbot. Especially colleagues who have little to do with the chatbot or study guidance in general because that way, you get a complete and representative picture. The team did several test runs and received much feedback, so the conversations could be adjusted and improved where necessary. Eric: “That we have done so much testing has to do with how our organisation is structured. Instructors may have different ways of doing things – for example, re-taking an exam may be free for one person and not for another. This all had to be factored into the conversations”.

After a soft launch was made within the student environment, the bot was further developed, and the big day finally arrived. The bot was now live and had a name and a face. Eric: “I can safely say the project has been a huge success! You can already see within the organisation that the chatbot is also being noticed within other departments, where colleagues left and right are wondering if a chatbot might be something for their department”.

Advise for other organisations

The implementation process was sometimes a bit of a ‘bumpy road’, but the team pulled through it together. Sarah Stoel, Knowledge and Quality Advisor has valuable tips for organisations considering having a chatbot: “Think carefully in advance about which questions you want the chatbot to include and pay a lot of attention to the analysis part so that you have a strong foundation from which you can continue to build on further. A choice you make in the beginning will certainly have an effect in the long term. You can, of course, make adjustments during the process, but many times that’s not very efficient”.

Sarah explains that you also need to keep each other in the loop and make agreements accordingly. “All chat conversations must be labelled, and everyone interprets the text differently. You must continue to seek each other out and communicate because you’ll be making important decisions based on this analysis.” Eric adds:

“The danger is that a fact can quickly become a suggestion. The assumption that is often made is that by answering a question, you also tackle the reason behind it, but the reason for that question sometimes lies somewhere else. A chatbot can do well by recognising the intention of a question, but the chatbot will give a literal answer to the actual question and not always to the underlying question. It’s good to be aware of this.”

Proud of the chatbot’s success

The team is very proud of what they have achieved. During the start-up phase, there was often doubt about the project and whether it was not too extensive and complex. But now there’s a chatbot in place that’s being well-used and helping students along the way. Eric:

“I am especially proud of how we have approached the process. We have kept each other sharp and critical regarding what we were doing. We felt the freedom to express any uncertainties if they arose. There were several times when we could have just decided to pull the plug, but instead, we managed to convince the organisation of the importance of continuing. As a team, we have truly taken responsibility for bringing this all to a successful conclusion.”

The team also sees lots of opportunities to expand the chatbot or implement new chatbots. In an ideal world, the chatbot would be the ‘go-to person’, where students could go with all their questions. Eric explains: “Now we present the bot as a digital study coach, but in the end, the chatbot can, of course, also become much more than that. It would be great if there were an app where the chatbot could answer any personal questions from students. If we could achieve that, the students would no longer even need to call or email. But that’s still just a pipe dream!”.