Two of Scandinavia’s leading financial institutions have very different strategies for their conversational AI-powered customer service but with one unifying element at the core: human augmentation
“We should go digital” is a common refrain in the financial services industry. Competition is increasingly happening online with smartphone apps, digital transactions and products that automate the customer experience. In the rush to avoid getting left behind, banks and insurance companies are looking to AI-driven virtual agents to smartly improve their limited customer service, with many even considering going all-digital on their conversational AI support efforts. But while the extensive use of digital technology has been a competitive edge across the industry, savvy financial institutions now know the truth: Going all-digital is a mistake.
Relying solely on automation to handle all customer service interactions is a slippery slope. Even the most advanced AI can, and will, stumble when tackling complex queries, spouting off responses that could potentially turn customers away, and negatively impact a brand.
Major European companies like Tryg and DNB have clearly understood that the path to finding the customer experience ‘sweet spot’ lies in using technology to augment existing human support staff. Both companies, for example, while implementing conversational AI technology in different ways, found that by leveraging the efficiency and predictability of AI against the expertise of their employees, they are able to achieve results without sacrificing the quality of service for which their customers have grown accustomed
Augmenting customer support with AI
Mitigating complexity was a key challenge for Tryg in providing security and peace-of-mind to its 4 million customers. Rather than using Artificial Intelligence as customer-facing first-line support, the insurance company built a virtual agent named RoSa designed to make internal operations more streamlined and efficient.
“When our outbound support teams speak to customers, they have all of the knowledge from the virtual agent instantly available and at their fingertips,” said Head of Process Excellence, Bjartmar Jensen at Tryg. “If the customer asks a question they don’t know the answer to, he or she can simply ask RoSa and get an answer immediately. In fact, 97 percent of the time RoSa has the right answer, or can provide useful relevant links, explain contract conditions, and provide guidance to internal processes.”
By making the company’s entire wealth of knowledge easily accessible to its employees, Tryg has ensured that none of its customers risk being greeted by support agents who are too inexperienced to handle their problems. Instead, using RoSa, even new employees can give customers expert advice and resolve inquiries according to best practices.
Since the virtual agent launched in July 2018, Tryg’s outbound customer support teams have been able to substantially resolve more cases and decrease the number of calls from support to back-office teams
Finding the right balance between human and machine
DNB is Norway’s largest bank, comparable in terms of assets to the top banks in the U.S. The bank chose to position its virtual agent, Aino, front-and-center in its digital customer-service strategy. The partially government-owned bank needed a solution to deal with thousands of daily inquiries to its website that, while easily handled by human support staff, still required considerable time and resources.
“The challenge for us is in handling the enormous amount of incoming traffic efficiently, while at the same time giving the best possible experience to our customers,” says Oyvind Brekke, EVP and Head of Digital Innovation at DNB. “Since launching our virtual agent in June 2018, we have already automated 50% of all incoming chat traffic.”
In a confident move by DNB, customers visiting the bank’s website are required to first interact with Aino who will attempt to solve their query on-the-spot. This takes care of the menial, repetitive tasks that would otherwise tie up support lines, freeing customer service staff to concentrate on more nuanced queries that the artificially intelligent virtual agent identifies and passes on.
To date, Aino has had over 1 million conversations with customers and can answer 3,500 different questions. Initially, the bank was seeing automation rates as high as 70%, however, DNB quickly discovered that there is such a thing as too much of a good thing. “The results are a lot higher than we expected in such a short time,” adds Brekke. “We work hard every day to make Aino even better and to find the right balance between human and machine.”
That balance is the key to Aino’s success. The bank dialed its capabilities back when the data made it clear that automation of around 50% of first-response queries (with high-quality automation) created the best customer experience. Overall, DNB now automates 17% of all its customer interactions across channels thanks to Aino, with that number expected to grow as its capabilities continue to expand.