Did you know that UK’s smart speaker adoption rate nearly tripled over the last two years? It’s estimated that at least 38% of the nation’s adults owned one by the start of 20211. What’s more, Siri alone now fields 25 billion requests each month across the Apple ecosystem2. Such statistics confirm our lived experience: Conversational AI is fast becoming part of everyday life.
For established financial institutions, this is a pressing challenge—and an unmissable opportunity.
The idea of managing finances through a conversation with a virtual assistant (VA) is far from new. But the normalisation of such conversational experiences is. Customers have grown comfortable chatting to Alexa or Google Assistant around the house. They’ve embraced conversational spend-tracking apps like Cleo. They’ve begun to not just expect conversational experiences, but to expect smooth, personalised ones.
The challenge for established banks is simple: how do you meet these high expectations? Many have self-service experiences that—built around a menu-based IVR and a basic chatbot—will inevitably disappoint a demanding digital native. And as switching providers becomes ever easier, there’s every chance they’ll look elsewhere: Research indicates that nothing influences as customer’s choice of bank and decision to stay loyal like its “ease and convenience of service”3.
But if the challenge is great, the opportunity is even greater. When financial services brands do respond to customer demand, and build effective experiences based on enterprise-grade conversational AI, it’s not just their CSAT scores that improve—it’s everything from their contact centre efficiency to their fraud prevention performance.
To understand why, we only need to look at the successful conversational AI strategies of trailblazing banks across the UK, Europe, and Australia…
Conversational AI at Commonwealth Bank of Australia
Commonwealth Bank of Australia’s (CBA) virtual assistant, ‘Ceba’, was recognised at the 2021 APAC Stevie® Awards—taking home Gold in the Innovative Use of Technology in Customer Service (Financial Services Industries) category. Its story shows just what’s possible when an initially fairly humble VA has access to advanced machine learning and conversational AI.
At the time of Ceba’s launch, typical customer enquiries included checking balances and activating cards. But over time, the bank has built is capabilities. Today, the VA can understand and answer around 70,000 different questions, and more than 90% of customer enquiries, in a matter of seconds. What’s more, it can take the right next step for the customer, directing customers to other self-service options, escalating the enquiry to a live agent, and even performing tasks itself.
Ceba proved especially valuable in 2020, as questions regarding coronavirus support helped to increase its workload—at times by as much as 1,000%. The VA fielded 581,000 conversations in March 2020 alone, with approximately four out of five relating to suspending or deferring payments. The VA’s flexibility was one of the keys to its success, allowing the bank to monitor the questions it received, and update answers in almost real time.
All in all, Ceba provides great example of conversational AI’s potential to impact customer experience and satisfaction. The NPS scores for CBA’s chat channel not only held consistent in 2020, but actually increased during some peak periods of the pandemic.
Conversation AI at Swedbank
Swedbank’s VA has also had a meteoric rise. It started out as an internal resource—a simple FAQ bot to help employees find information faster. Soon, however, it had evolved to live in the leading Swedish bank’s log-in environment and support its customers directly.
One key evolutionary leap has been the integration of Swedbank’s VA with its CRM, which gives the VA insight into everything from the webpages a customer has tried to access, to the cards associated with their account. As Hans Lindhom, Team Manager at Swedbank, told attendees at our webinar, “The integration with our CRM helps customers complete specific act flows like replacing their card and ordering historical account statements, all without interacting with an agent. It’s also helping us provide more personalized answers.”
It’s easy to underestimate the impact that automating such everyday tasks can have on contact centre efficiency. In Swedbank’s case, its VA now answers 80% of customer questions and handles 25% of all debit card replacements, freeing up agent time.
Conversational AI at BNP Paribus
BNP Paribas Personal Finance Spain launched its VA with efficiency as a key objective. “With our virtual assistant project, we seek to achieve the most important improvements for any customer service model: greater customer satisfaction and increased efficiency,” says Maria Paz Puchol, Director of Operational Digital Transformation.
For Mari, bringing the right people onto the project was key. “One of the success factors was the involvement of members of the BNP Paribas Personal Finance team in Spain from different levels of the business from the beginning,” she explains. “This way, we were able to define what it was that we needed more easily and more accurately.” And what success the bank has seen—its VA has a 96% accuracy in understanding queries, and the volume of calls being managed by its contact centre agents has decreased by 46%.
Every great banking VA has a robust strategy for handing difficult or sensitive interactions to human agents. The VA at BNP Paribas can seamlessly transfer a customer to a contact centre agent if it discovers it can’t help or if it detects that the customer is frustrated.
But many financial institutions aren’t just using AI to deliver conversational banking experiences. They’re using it to fight the rising tide of fraud.
Secure conversational banking at NatWest
In recent years, criminals have become increasingly adept at circumventing traditional authentication processes, whether by purchasing customers’ personally identifiable information on the dark web, using malware to steal credentials and one-time passwords from customers’ devices, performing SIM-swap attacks, or socially engineering contact centre agents.
One solution is to build AI-powered biometrics into your conversational banking experiences, and authenticate customers based on their inherent characteristics, such as the sound of their voice and the way they swipe, type, and hold their device.
Voice biometrics technology is an increasingly popular choice for banks, because it can simplify and accelerate service for genuine customers, while equipping fraud prevention teams to identify known criminals within seconds of them placing a call.
One bank that’s making great use of voice biometrics is NatWest Group. As Jason Costain, Head of Fraud, explains, “It’s not just about stopping financial loss—it’s about disrupting criminals. For example, one prolific fraudster we identified was connected to suspect logins on 1,500 bank accounts. That’s helped us protect potential fraud victims and identify the ‘mules’ being used by the crime network to perpetrate fraud, leading to two arrests so far.”
In its first year using the biometrics solution, NatWest screened over 17 million calls, raised 23,000 fraud alerts—and saw 300% ROI.
Conversational AI in financial services: Learning from the pioneers
As financial institutions look to improve their self-service experiences and deliver the kind of effortless, conversational interactions that consumer electronics are teaching us all to expect, they can take encouragement and inspiration from the stories shared above. From seamless transitions to live agents at BNP Paribus, to monitoring incoming questions at CBA, they demonstrate many of the best practices that are essential to making conversational banking experiences a runaway success.