Have you used the customer service app with your bank the past year, or received an unexpected email with an offer you were actually interested in? Maybe it was a well-timed mortgage re-fi or even some savings at a favorite store.
There’s an enduring and unfortunate misperception that AI serves only to replace human workers and irritate human clients. But what if we don’t have to be locked in a zero-sum game with the growing legion of digital intelligence?
An increasing number of banks and insurers are finding that it’s almost impossible to meet the rising customer expectations and needs that digital services and apps have unleashed. It was an acute problem for financial institutions even before COVID-19 shuttered branches and agencies and made social distancing as important as remembering the last four digits of your social.
Ankit Chhajer, head of artificial intelligence at UK-based NatWest Group, saw the impact firsthand as the pandemic set in last spring.
“Our lines were getting 2,000 calls a day, and suddenly they were taking 20,000 calls,” Chhajer said. “It really left banks and companies with no other choice than to leverage conversational AI.”
NatWest and its previous incarnation, RBS Group, has for years been working with IBM to develop virtual human solutions for its customer service needs in both the real and digital realms. As the bank confronted the challenges of COVID-19, the companies deepened their collaboration to, in Chhajer’s words, “supercharge” their AI.
“It’s pretty cool, these questions we’re working on around how to empower people with AI,” said Jean-Phillippe Desbiolles, global VP for data, cognitive and AI for Financial Services at IBM. “It’s how to engage the customer with what we’re calling the augmented banker and the augmented insurer.”
As customers shifted to online banking during COVID-19, AI helped manage the resulting surge of customer service requests.
Desbiolles works closely with the world’s largest banks and insurers to incorporate AI into their day-to-day workflows. And while efficiency and cost savings are important, one thing remains paramount to building good AI: humanity. That’s humanity in terms of putting customers first as well as infusing the output with compassion and humility. And then there’s the matter of the deep collaboration required to produce customized AI tools.
Banks and insurers put AI to work
That’s why Desbiolles and his team set up what he calls a “cognitive factory” with financial services clients. These factories are an AI-focused business unit that, with their human-centered approach, aren’t solely staffed by engineers and data scientists but also by psychologists, linguists and specialists from other disciplines. It takes a diverse array of skills to ultimately design, build, train and deploy AI at scale in order to achieve a level of performance that meets the needs of customers and employees.
“We are shaping innovative use cases and designing these machine learning models, and these systems are learning from whom? From us humans,” Desbiolles points out. “We need the right balance between hard and soft skills if we want to ensure that the outcome of AI models can be trusted by the end user.”
Building that trust is a matter of transparency, relatability and robustness, Desbiolles acknowledges, but at the end of the day, it has to be the result of the collaboration between people and programs.
Consumers now prefer a mix of physical and digital banking, which AI can help manage.
To bring this to life, his cognitive factories have developed AI-powered tools that include consumer-facing ones (like those chatbots, virtual assistants and apps) as well as those used by middle- and back-office employees. While the former tend to garner the most buzz, the latter are no less important for improving the human experience in a complex industry like financial services.
For example, AI equipped with natural language processing can review and produce insights from analyst reports in a fraction of the time it would take a know-your-customer specialist; this frees the KYC specialist to focus on critical onboarding decisions for clients. In another case, for bankers working with consumers, AI-powered research systems can significantly enhance their access to key information about various products the bank has available, empowering the bankers to deliver better service.
“The best experts cannot be everywhere in every branch,” Laurent Prud’hon, the cognitive factory leader at the French bank Credit Mutuel, said. “It’s a way to take the knowledge of the best experts and to spread it and to deploy it everywhere in all conversations inside the company.”
For AI to augment the human experience, however, it must also be user-friendly. At Credit Mutuel, for instance, bankers obtain information by typing conversational questions into a chat with a virtual assistant, as opposed to the string of words more typically associated with using a search engine.
Learning from AI, and training it
The ability to understand and respond to human language is even more important for chatbots that serve banking customers directly.
One bank found that its chatbots, which were managed by IBM Watson, successfully answered 55% of all customer questions and requests—impressive numbers that allow for the other 45% to be quickly referred to human bankers. Part of effectively implementing AI is determining when it’s time for the AI to pass the baton.
“It’s about finding the right place for AI within the overall client journey,” Desbiolles said.
AI can also help banking and insurance employees deliver better service by streamlining financial information and products.
The human minds behind AI, meanwhile, are on a journey of their own.
AI must be trained continuously to keep up with new data and trends. Developers and customer representatives remain vigilant in improving and refining the technology, including efforts to identify and mitigate problems such as bias.
After all, nothing should get in the way of the customer experience, even in the age of AI. “If you don’t put humans at the heart of everything we think and we do,” Desbiolles said, “we are missing something big.”