Aggrieved clients often require immediate professional support from call centre service providers when they reach out. Unfortunately, they frequently endure long waiting times before connecting with an agent.
 
This protracted delay has a detrimental effect on customer experience, impeding growth and diminishing the Net Promoter Score (NPS), which gauges customer loyalty. Moreover, customers commonly voice various grievances, such as inadequate network service, subpar customer service, and affordability issues.
 

However, this situation may soon be a thing of the past, thanks to a team of innovative masters students at the University of the Western Cape (UWC), who have developed an Artificial Intelligence (AI)-based solution to enhance customer experience and retention within call centres.
 
The tool, created by Takudzwa Sikumbuzo Mzembegwa, Thandile Nododile, and Sange Mafinyongwana, all pursuing masters degrees in the Department of Computer Sciences at UWC, has not only garnered significant industry interest but has also earned them the coveted inaugural Southern Africa Telecommunication Network and Applications Conference (SATNAC) 2023 Industry Solutions Challenge.

SATNAC celebrated its 25th anniversary with a call to postgraduate and undergraduate students to participate in one of four industry-inspired challenges issued by its top sponsors. The four categories were sponsored by ZTE, Openserve, Huawei, and AMDOCS – UWC students formed several teams, which competed for the challenges independently.

The UWC trio was one of those teams and emerged victorious in one of the four challenges. In recognition of their achievement, the team has been offered a six-month internship by AMDOCS, promising hands-on skill enhancement and exposure to global leaders, amongst other benefits.
 

"This opportunity represents a gateway for us to engage and collaborate with prominent figures in the industry, " the trio expressed in an interview. “It allows us to apply the knowledge imparted through lectures and the mentorship of our supervisor, Prof Clement Nyirenda, during our academic tenure. As UWC masters students, we take pride in elevating our university's reputation and showcasing the values and expertise we have acquired.”
 
The students' solution entails an automated query resolution system featuring skill-based routing for call centres. Leveraging natural language processing (NLP) and semantic text similarities, intent recognition is employed to connect the customer with a skilled agent if a customer's query remains unresolved.
 
"The primary issue in call centres is extended wait times and redirection to agents without the necessary skills, resulting in users being shifted between agents until they find one capable of assisting them," the students explained.
 
To address this problem, their implementation harnesses Artificial Intelligence. It involves NLP, enabling users to express their queries. NLP then vectorises user input and performs semantic text similarity matching with a repository of vectorised frequently asked questions (FAQs).
 
 If a match is identified, the user is provided with an existing solution. In cases where no similar questions exist, a ChatGPT API, with context stored in Pinecone VectorDB, is employed. If the user's issue remains unresolved by the automated process, they are directed to a suitably skilled agent. "This approach minimises call centre wait times, optimises agent availability, enhances user retention, and reduces call duration as users are swiftly connected to proficient agents, avoiding unnecessary transfers."
 

Professor Clement Nyirenda, Director of eResearch in the Office of the Deputy Vice-Chancellor: Research & Innovation, who serves as the MSc supervisor for all the students in the UWC winning team, congratulated them and said they competed independently as he was not involved in their work at all.

He said their win speaks volumes about the work done at UWC.  “This win shows that the graduates we produce in the Department of Computer Science can compete with the best brains in the land. All three members of this group are home-bred; they all started with us from their first year. The quality of their work reflects the input from all the UWC lecturers who taught or mentored them over the years. The fact that the AMDOCS project is, in no way, related to their MSc projects shows that our graduates can think out of the box when faced with real-world problems. This really gives us confidence in the work we have been doing over the years and inspires us to do even more.”

Meet the Visionary Minds Behind the Initiative

Takudzwa Sikumbuzo Mzembegwa

Takudzwa is an MSc Computer Science student specializing in real-time pipe burst detection using AI. He is a MasterCard Foundation Scholar, achieving cum laude for both his BSc and BSc Honours in Computer Science. He currently serves as the vice-president of the UWC IT Society and actively collaborates on web and mobile application development projects with faculty and PhD students in the Department of Dentistry at UWC.

Thandile Nododile

Hailing from Umtata in the Eastern Cape, Thandile is pursuing a masters degree in Computer Science at the University of the Western Cape, where his research revolves around the exciting realms of blockchain technology and deep reinforcement learning. “By day, I'm a seasoned senior blockchain developer, and by night, I'm fuelled by my passion for entrepreneurship and Bitcoin.”

Sange Mafinyongwana

Pursuing a masters in Computer Science at the University of Western Cape researching Enhancing Microservice API Security using AI.  He also holds the position of president of the UWC IT Society, where he fosters a community of technology enthusiasts, providing opportunities for growth and learning. He is also the co-founder of tech start-up, LenoKids (Pty) Ltd, blending academic prowess with entrepreneurial spirit.