One way digitally-forward banks are disrupting traditional large banks and posing a real threat is through adopting a conversational banking strategy, turning what has previously been a big challenge into a relatively significant advantage. While bot-building big banks are stuck with first-gen legacy “chatbot” systems and all their urgent problems, by contrast, community and regional banks are just now starting to upgrade their technology and seeing real results.
Regional Norwegian bank, Sparebank 1 SR-Bank, was one of the first in the country to go all-in on a conversational AI-powered virtual agent. A move that Lead Technological Strategist, Ramtin Matin, says has paid off: “Since launching our virtual banking agent, Banki, in 2016, we are now saving the equivalent of 30 full-time employees based on our current volume of traffic.” He adds, “Only 25% of interactions are escalated to human support, the majority of which are for complex processes that the AI can’t handle yet.”
Institutions like SR-Bank are partnering with providers to deploy second-generation AI-powered conversational agents, with pre-trained industry-specific data, aimed to achieve a high level of accuracy, even from a cold start. These virtual agents have the human-like ability to teach themselves new tricks and streamline frontline interactions.
They use AI to understand customer requests, engage in real give-and-take interactions, and accurately interpret them into actionable outcomes. It’s now possible, with conversational AI, for virtual agents to perform relatively complex tasks on behalf of customers (e.g. bill pay, transfer funds, open accounts, apply for credit card and loans), identify possible fraud and offer advice and recommendations about specific financial products and services.
The idea is that rather than spending precious (human) resources on tackling thousands of daily requests for relatively easy-yet-time-intensive tasks, a bank’s customer support staff can focus on more delicate, nuanced cases. A virtual agent then acts as a kind of “the first responder,” automating the bulk of interactions and forwarding anything outside of its remit to skilled human support.
For example, banks like DNB and Nordea are already leveraging conversational banking technology. They have raised the bar by mining vast amounts of data while providing highly-targeted and personalized customer interactions.
“When we were initially looking at automating our online banking chat, we anticipated a 20-30% automation rate,” says Jan Thomas Lerstein, SVP IT Emerging Technologies at Norway’s largest bank, DNB. “What we achieved instead was an automation rate of 70%, essentially overnight, simply by placing our virtual agent as front-line support.”
This result was beyond DNB’s expectations; however, the bank quickly realized it would need to find a balance between automation and customer satisfaction eventually settling on automating 51% of all frontline queries and text-based customer support tasks, with the remaining 49% accurately handed off to the appropriate human advisors. Since launching in June 2018, DNB’s virtual agent, Aino now automates 20% of DNB’s total customer inquiries across all channels, including phone and email, and can answer over 3,500 product and service-related questions.
The adoption of conversational AI as a primary channel for support has had a positive reaction with customers, as well. “We find that, overwhelmingly, our customers are happy to speak with Banki,” says Matin, who is adamant about keeping customers in the loop when developing new functionality for SR-Bank’s virtual agent. “Keeping lines of communication open with the users and customers actually using the technology has meant we see an upward trend in positive feedback towards the 60% mark.”