What do you do when your contact center solution isn’t really solving your problems? For NovaKBM, this wasn’t just a question - it was a reality. The simple chat feature they implemented to automate their customer service requests continued to be uncontained and led to an overwhelming need for help in their contact centers.
As one of Slovenia's leading banks, NovaKBM offers banking services from private banking to corporate finance and prides themselves on providing incredible customer service experiences. With the majority of their chats still being transferred to live agents, providing that level of service became increasingly more difficult.
With an unwavering commitment to meeting customer needs, Aleksandra Brdar Turk and her team at NovaKBM's Banking Operations set out on a mission to create the perfect digital banking assistant for their members.
The Need to Contain Requests with Conversational AI for Customer Service
Established in 1862, NovaKBM was in the midst of a digital transformation with the goal of becoming the most digitally advanced call center across the entire financial industry within the country. With AI automation revolutionizing the contact center scene, now was the perfect time to implement the change to self-service automation.
After analyzing their customer interactions, NovaKBM realized that many phone calls handled the same repetitive questions and experienced similar types of support requests from their members. They also realized that there was an opportunity to enhance customer satisfaction by offering additional self-service options that would streamline processes for both members and business operations.
Meeting the Language Challenge with the Multilingual Virtual Assistant
Choosing the right contact center solution is difficult in itself but NovaKBM found themselves with another problem. In a country with roughly 2 million people, the struggle became not just finding a contact center solution that was robust enough to increase their containment rate - it was also finding a platform that could build a Slovenian speaking virtual assistant. Without NLU support in the country’s native language, automation would be impossible.
“Being a small country of only 2 million people, we are very proud of our culture and our language.” says Aleksandra Brdar Turk, Director Banking Operations, Nova KBM. “ It would be inconceivable to have a chatbot that couldn’t speak Slovenian to a person. So that was a must. It was like our basic requirement.”
Finding a platform that would be readily available and able to support a lesser-known language was difficult. Many of the solutions they came across were either too expensive or were not capable of customizing an NLU to adapt their nation's language. As a language, Slovenian also had a few difficulties. It is only spoken by roughly 2.5 million people globally and has intricate grammatical structures that are shared exclusively by two languages worldwide: the other one being Latin.
Determined to find a solution, Aleksandra and her team explored possibilities across Europe - looking through demos and testing different platforms - unfortunately, none offered Slovenian language support for conversational AI.
That is, until they found Kore.ai and its multilingual virtual assistant.
The Solution
From the beginning, Kore.ai embraced NovaKBM's most critical requirement - natural language understanding (NLU) for Slovenian and immediately rose to the challenge. With a dedicated language development team at Kore.ai, Aleksandra and her team had the language support they needed to accomplish their most crucial milestone. In just 6 weeks, Kore.ai was able to completely build an NLU that supported the Slovenian language offering an incredible time to value.
When using the platform itself, Aleksandra and her team found the tool to be easy to use and were pleasantly surprised that they didn’t need to pull in developers to assist with building the virtual assistant. Users and testers who contributed to the platform at NovaKBM, found the user interface incredibly easy to use, allowing them to begin creating content without any prior training.
The results are incredible. In just 6 months after deploying the assistant into their Nova KBM mobile app, chats leapt from 5000 a month to over 50,000 per month. Even with this astonishing increase in capacity, NovaKBM found that live agents went from servicing 24% of chat messages to just 6%. With this incredible change, the company quickly looked into client satisfaction using thorough evaluations to measure their NPS scores. After implementing a virtual assistant, NovaKBM saw an impressive 25% increase in these ratings over a short period of time!
Today, NovaKBM is operating at 94% automation and are using the time saved to train their live agents on improving their skills, widening their knowledge and learning to handle difficult customer situations. With this incredible success, it’s one thing to read about it and another to hear the enthusiasm from Aleksandra herself. We had a chance to listen to Aleksandra speak about the project with VUX World - the call is inspiring. We highly recommend tuning into their webinar to experience the inspirational message first-hand.
