Increases productivity and saves more than 100 hours a month per customer service employee
Increases capacity to handle ~2,500 queries a month
Enhances accuracy above 80% for answerable queries
Nanyang Technological University, Singapore (NTU Singapore) sought more advanced generative AI solutions to further enhance its chatbot, with Google Cloud solutions being the key to unlocking a much-improved experience for students and staff.
Nanyang Technological University, Singapore (NTU Singapore) sought more advanced generative AI solutions to further enhance its chatbot, with Google Cloud solutions being the key to unlocking a much-improved experience for students and staff.
The tertiary school system can be complicated to navigate, not just for staff but also for students who are coming to terms with more responsibilities and potential future pathways. For Nanyang Technological University, Singapore (NTU Singapore), over 33,000 undergraduate and postgraduate students are in its current programmes, and for a selection of them, the campus is also home.
That means that the organization has to be ready and equipped to handle any and all queries that might arise during the students’ academic pursuits. Having recognized the importance of technology and how it can make life easier for all, NTU Singapore began its digital transformation seven years ago, with cloud infrastructure a critical component.
As an extension of the migration, the Enterprise chatbot team, led by Deputy Director Ng Kee Haur, also began looking at creating chatbots as additive assistance for various working processes. The first Lyon chatbot built for freshmen was an intent-based solution, often requiring the backend team to spend long hours and intense effort to cover all possible grounds.
The technological constraint and human effort required to update intents, classify queries, curtailed its effectiveness. This led the team to seek more advanced generative AI solutions and, eventually, to Google Cloud partner CloudMile.
Together, a conversational flow was developed for a next-generation chatbot — starting with a Lyon Housing chatbot. Key steps such as intent detection, query rewriting, persona classification, and response generation were implemented. Leveraging the capabilities of Dialog Flow CX and Gemini to improve query coverage and engagement quality, a new and enhanced experience became feasible.
We approached different solution providers, trying to understand the framework, the capability of their models, and how easy or difficult it is to implement based on our requirements. It was truly a learning process before arriving at our decision.
Ng Kee Haur
Deputy Director, Enterprise IT, NTU Singapore
Transforming the student housing experience, revolutionizing staff support
The highly collaborative and customer-centric approach to developing the Lyon Housing chatbot created opportunities for NTU Singapore to further modernize its approach to serving the students and helping staff.
With the implementation of Dialog Flow CX, Gemini Large Language Models (LLMs), the immense amount of data collected previously and from new queries can be properly cleansed and evaluated. This produces actionable insights for the team to explore further ways to improve the chatbot.
Jacquelynn Koko, Senior Assistant Manager, cited the example of numerous queries asking about bed sizes in the dormitories, an element of the campus housing experience that surfaced with the Lyon Housing chatbot. This meant that NTU Singapore could now account for this growing concern, and address it for future students.
We have had a successful launch for housing. We are also working on admissions. There are a lot more topics that we need to explore in a student's life cycle. All these various topics also mean opportunities for us to utilize genAI to serve our customers.
Ng Kee Haur
Deputy Director, NTU Singapore, Enterprise IT
The performance of the Lyon Housing chatbot has also demonstrated the value of having AI do the heavy lifting when it comes to commonly asked questions. Overall, the workforce usually in charge of attending to housing queries are saving more than 100 hours per month, with more time to focus on more complex issues that require the human touch.
With a scalable capacity currently handling an average of 2,500 queries monthly, the Lyon Housing chatbot continues to perform reliably and supports the population through peak periods and slower months. The chatbot can fetch answers to more than 80% of queries, while sending those complex cases that cannot be answered to human agents.
Having seen the success and possibilities of genAI solutions with the Lyon Housing chatbot, the university continues to look to other areas where such implementation can be carried out.
“We have had a successful launch for housing. We are also working on admissions. There are a lot more topics that we need to explore in a student's life cycle,” says Ng. “All these various topics also mean opportunities for us to utilize genAI to serve our customers.”
A research-intensive public university, Nanyang Technological University, Singapore (NTU Singapore) has 33,000 undergraduate and postgraduate students in the Engineering, Business, Science, Medicine, Humanities, Arts, Social Sciences, and graduate colleges.
Industry: Education
Location: Singapore
Products: Dialogflow CX, Gemini
About Google Cloud partner — CloudMile
CloudMile, a leading AI and cloud service provider in Asia, is focused on helping corporate clients implement AI to drive growth, leveraging machine learning and big data analysis to assist businesses with forecasting and industrial upgrades.
