How can innovative AI-powered tools optimise customer service in financial services?

A Netflix-plus offering is within reach, enabling hyper-personalisation and a democratisation of money management advice, but collaborating with trusted partners is essential, according to experts
Artificial Intelligence Touch Screen

The introduction of generative AI has turned the head of many business leaders, eliciting mixed reactions. So, while already-high consumer expectations have been further elevated by recent advancements, how could – and should – financial service operators approach AI-powered customer service without turning people off?

Sol Enenmoh, director of digitalisation at HSBC, acknowledges that it’s a tricky balance to strike, but is thrilled by the confluence of technologies – including generative AI, quantum computing, and cloud computing – that promises to evolve digital experiences for customers and also employees. “We are in a position where things are shifting from theory into practice very quickly,” he enthuses. “AI and interconnectivity present incredible opportunities to offer hyper-personalisation preemptively.”

AI can deliver personalisation and enterprise search solutions

Amazon and Netflix signal the direction of travel for customer service in other industries, posits Sami Helin, managing director EMEA of Coveo, a global software-as-a-service firm with an AI platform that delivers personalisation and enterprise search solutions. “They have figured out that the key to success is to serve each individual personally rather than take a one-size-fits-all approach that will always miss the mark.”

Helin says capabilities similar to Netflix’s “what to watch next” suggestion prompts are now possible in financial services. “You can exceed expectations and provide a tailored experience. It is not rocket science, but to make those recommendations happen at scale in milliseconds is difficult. Here machine learning and AI are extremely necessary.” 

Coveo’s “mission”, Helin continues, is to harness AI to “provide relevant, personalised and profitable experiences for people across employee productivity, customer service, or ecommerce.” 

Getting sentimental and reading the digital room

Enenmoh argues that AI is now proficient enough to “take an extra step” on from the Netflix example by going deeper to understand the “mind state” of the user or customer. “AI presents an opportunity to serve you better, having learnt, pivoted and grown with you organically and knowing your current state of mind,” he explains. “That’s the power it has.” 

With such competence, Enenmoh says, a financial services operator will know – purely based on the sentiment analysis – whether to offer additional services, if the customer is in a good mood. The AI can read the digital room and act accordingly.

This development area also delights Reena Sukha, chief information officer at Investec, which uses AI on voice calls to understand the emotional state of customers. AI is beginning to “humanise the digital client experience”, and, in tandem, the general public is becoming more comfortable with the ubiquitous technology. “People don’t even know the difference anymore,” she says, nodding to Alan Turing’s famous test for machine intelligence, proposed in 1950. 

Originally called the imitation game, the Turing test was designed to gauge the cleverness of a machine compared to humans. Essentially, if a machine displays intelligent behaviour equivalent to, or indistinguishable from, that of a human, it passes the Turing test.

In the context of customer service, all that matters is the quality of experience provided by either man, machine, or a combination of both, states Sukha. She quotes the late American author Maya Angelou. “At the end of the day people won’t remember what you said or did, they will remember how you made them feel.”

That line “really resonates” with the Investec CIO. “I’m excited at how AI is transforming how we interact with our clients,” she says. “The more it is becoming humanised, the more comfortable we – the humans – will be with the technology.”

Co-creating better customer experience using deeper analysis

Alongside the smarter, AI-driven relationships with customers, Investec’s sentiment data also inform and enhance those interactions for everyone. In addition, investment in AI voice technology for customer calls has proved valuable in various ways. Initially, Investec rolled out this solution as it received a high volume of calls from clients, which was costly and the team saw opportunities to improve the customer experience. 

A “system analysis” helped the customer service team identify the top 10 reasons for the calls. “We used that information to determine the roadmap for what we were going to do with our platform,” says Sukha. For example, many callers couldn’t find functions online, meaning better signposting was required. Notably, there were activities customers felt “safer” doing with a human rather than online or via a banking app – changing bank account details, for example.

A deeper investigation, through interviewing clients, allowed Investec to test the level of comfort with AI while educating the would-be callers to put them at ease and encourage them to use online functions. “We asked them whether they would feel comfortable with certain things if we were to introduce them,” Sukha continues. 

The exercise was a win-win scenario; it engaged Investec’s clients, understood their pain points, and enabled a co-creation of suitable digital tools to solve those challenges. And AI is assisting Investec’s callers in another way: voice biometrics makes authenticating users quicker than before, as well as understanding their emotional state, which has improved efficiencies and seen the length of interactions drop by around 20%.

Chris Waring, head of digital journeys at NatWest, agrees that assessing how and with whom customers wish to communicate, depending on their needs, is a fascinating and rapidly evolving area. 

Chatbots learning new languages and ways to communicate

NatWest’s AI-powered digital assistant, Cora, has developed pleasingly since its inception in 2018 when it was more of a “frequently-asked-questions” platform. The chatbot is always available and can support customers with day-to-day banking queries. Last year, there were 10.4 million conversations throughout the year with Cora, with almost half (48%) requiring no human input.

“Being able to surface those service journeys through the channel of choice – whether that’s Cora, or via the website, or through WhatsApp – is what we are trying to drive forward,” says Waring. “While Cora started helping customers with basic queries, it’s increasingly been integrated into various servicing journeys across the bank.” For instance, NatWest has begun using the chatbot to initiate loan deals.

 Waring concedes Cora is still not the finished article, but says NatWest is “continually improving” the product and focussing on customer personalisation. However, he stresses that relying on chatbots and conversational AI is not necessarily appropriate for “complex requests that are better served in different channels”.

Geoff Branch, enterprise account executive at Coveo, is equally ambivalent about chatbots, new and old. “When the first chatbots were released, everyone thought they would be the answer to everything, the silver bullet,” he says. “But chatbots – and now in the case of generative AI, like ChatGPT – are a good example where great technology can help you without human intervention for about 80% of searches.” 

What happens the remainder of the time, though? Branch answers: “We need that augmented service for high-value human interactions.” He adds that generative AI is “going to be right for some journeys, and, as is the case in the regulated environment, not for every journey.”

He observes that AI has changed the way people search for information by enabling more conversational and personalised interactions with search engines and other online resources – instead of simply typing in a keyword or phrase and expecting a list of relevant results, people can ask more natural language questions and receive more targeted and specific answers.

“This shift has led to a change in how people approach information seeking,” Branch surmises. Rather than just looking for a quick answer to a specific question, people are now seeking more in-depth advice and guidance on a variety of topics. For example, instead of searching for ‘how to make spaghetti,’  a person might now ask ‘what are some easy spaghetti recipes for beginners?’”

Interestingly, many predict that “robo-advisors” – part human, part AI – hybrid solutions will soon help customers with money management, including Waring.

Narrowing the financial advice gap with hybrid solutions

Royal London estimates that 39 million UK adults – of roughly 52 million in total – have fallen into the so-called “advice gap”, meaning approximately 75% don’t take any form of professional advice or guidance regarding their finances.

Given that Brits under 25 are more likely to turn to social media for financial advice than a professional adviser, according to Open Money research, there is a huge opportunity to innovate in this space for financial service operators and better engage present and future customers.
There are challenges when moving from person-to-person interactions to a digital setting to ensure we always translate that in a way that is compliant with regulation, even for edge cases, where a particular request is unusual or doesn’t fit the standard or anticipated pattern, says Waring. In some cases, it can be an improvement, as you can often explain options and product features more clearly in a digital setting.

“In commercial banking,” says Waring, “we could use AI and customer data to help structure financing appropriate for a company’s ambitions, together with relationship managers. We’re seeing a blend emerging between customers that want to speak predominantly to a relationship manager at a bank, in a traditional fashion, versus other customers that genuinely don’t want to start the journey that way.

“Giving customers the flexibility and option to bring in some relationship management debt advice, and so on, with digital offering is an interesting prospect. But, as banks, we must evolve to deal with that and bring people in at the right moment.”

Phil Williams, group chief operating officer at The Clear Group, states that “trust is the key thing here,” and financial services operators that update their incentive-based advice model would more likely capture the custom of people conditioned to think financial professionals are either for the wealthy or not worth the cost. 

Many of us are going online for advice, with limited success. “We’ve been taught as consumers that high-quality advice costs money, but the person who has the most vested interest in the outcome for me is me,” Williams says. “There is a sense that people will do their own research, go on community forums and social media. AI, though, can democratise financial advice. Providing the trust element is there – and I’m not sure it is at the moment – then high-quality advice could be available to everyone.”

Data-sharing and democratising high-quality money management

Business leaders may also lack trust in AI, not least because purportedly game-changing solutions have deluged the market. “There is a little bit of snake-oil salesmanship, something genius, some hype, and something dubious in what is being offered right now,” says Sandeep Dubbireddi, innovation consulting senior director at Salesforce.

However, shifting the value proposition to make AI products with trust and reliability in mind, including guardrails intended to help guide customers make ethically-informed choices, chimes with him. 

“Generative AI use cases are already taking flight across many industries. In wealth management, human advisors beat fintech solutions today, even those narrowly focused on specific asset classes and strategies, because humans are heavily influenced by idiosyncratic hopes, dreams, and fears,” Dubbireddi says. 

“LLMs used in Generative AI provide a tidy solution to these problems with a better understanding and thus a better navigation of consumers’ financial decisions. These systems can answer questions, evaluate tradeoffs, and ultimately factor human context into decision making.” 

Another thorny topic is personal data, on which AI programmes in financial services, as elsewhere, rely. There is a “dichotomy” here, says Coveo’s Helin. “On one hand, many people want total personalisation for relevant things, but those same people don’t want to share any details,” he suggests. “The question is: where is the sweet spot?” Helin adds that for most customers, unsolicited communications can come across as spammy, if not “creepy”, and urges organisations to tread carefully.

Sukha offers an elegant approach to “take away the creepiness” for financial service operators: be transparent with customers about the data gathered, how they could profit in terms of tailored products and services, and give them control, so they can opt in or out of certain things. 

Whether granted permission or not, there are other critical advantages of using AI-driven analysis, states Williams. “Having the data allows us to identify vulnerable customers,” he says. “Using sentiment analysis, for example, to understand whether someone is financially distressed, would be an amazing benefit – not to sell them more products but guide them through the journey.”

Education, experimenting and ecosystems – how to improve customer service

There is no doubt AI is improving customer service in the financial services industry, but there is more that operators can do to evolve the offering further. The lack of genuine knowledge at c-suite level is holding back progress, says Enenmoh, and education is crucial so that AI projects – even if pilots – will gain sponsorship. 

Further, HSBC’s director of digitalisation stresses the importance of developing and nurturing an ecosystem of trusted expert partners. “The notion that you can be successful by building everything yourself is shortsighted,” he says. “Being in an active and collaborative ecosystem, leveraging existing know-how from future-facing companies, allows you to be more expansive with your horizons.”

Salesforce’s Dubbireddi builds on this theme. “Systems relying on generative AI evolve their behaviour in response to the data they encounter, as well as the history of the interactions they experience. This demands continuous evolution. It’s impossible to keep up with the pace of innovation, so partner with individuals and companies that have a continuing commitment to your company’s success,” he says.

Finally, Sukha warns financial services operators not to be blinded by AI and to never lose sight of the humans the industry serves. “Ultimately, it’s not going to solve all your problems, and it is a tool to help customers,” she says. “So use it for real and meaningful problems.”