Home > What’s UP > blog > Machine learning in digital banking: The rise of the bots

blog

02. 01. 2018

Machine learning in digital banking: The rise of the bots

Now is the time for banks to embrace artificial intelligence and machine learning in digital banking. Their customers and bottomline will thank them for it later.

Artificial intelligence (AI) and machine learning (ML) are not going anywhere in financial services. Far from it: they keep popping up in industry outlooks and predictions as one of the must-focus areas in 2018 and beyond. But what are AI and ML to begin with? They are computing systems that can engage in human-like thought processes such as learning, reasoning and self-correction, as defined by Marketforce.

Let’s see a few examples. Customers can interact with AI-powered virtual assistants and chatbots, and complex algorithms can go through huge amounts of data behind the scenes to generate better customer insights that banks can cash in on. And that’s not all: banks could also make great use of artificial intelligence in back office operations, compliance, customer experience, product delivery, risk management and marketing.

It’s a long list for such short words, and that’s exactly why AI and ML shouldn’t be neglected.  And there’s still more. AI-driven data analytics can help banks understand their customers better and personalize products to satisfy their exact needs. Or as The Financial Brand has put it, “we need to look for ways to communicate to an ‘audience of one,’ using artificial intelligence systems that constantly work in the background to enhance every step of the customer journey.”

All eyes on AI-driven analytics

Recent survey findings underscore the emergence of AI as a must-have business tool for banks. About 51% of respondents in Marketforce’s new retail banking report say that their organization will be making some serious investment in AI-driven analytics in the next three years. And within just five years, most respondents expect to use AI-driven analytics to help customers self-serve, improve pricing and personalization, and even to come up with new products.

The digital consumer is no stranger to artificial intelligence thanks to tech and e-commerce giants that are fast becoming masters of everything AI (think Google, Amazon, Facebook and Apple, or simply GAFA). And clients more and more expect their service providers to know them, the Digital Banking Report says. They’ve got used to GAFA’s automated assistants, such as Google Home, Apple’s Siri and Amazon’s Alexa, so it’s time banks also jumped on these new channels, PwC suggests.

The case for artificial intelligence

But it’s not only the GAFA group that banks could learn from: other retailers are also making good use of AI and ML systems to boost sales. Here are some examples, as compiled by BizTech magazine:

  • TGI Friday’s had collected data from point of sale systems, social media posts, credit card transactions and mobile devices, then used AI, bots and mobile apps to develop tailored ad campaigns. The technology was first put to the test last summer, during a heatwave in New York. If an app user, say, tweeted that they were craving a Long Island Iced Tea, TGI Friday’s was straight on it. The bots flagged the post and the algorithms quickly matched it to the customer in the restaurant’s database. TGI Friday’s then sent a message to the person’s mobile, referencing the drink and inviting them to pop into a participating restaurant during happy hour. The result? Revenue soared by 70% during happy hour.
  • Coca-Cola used the data it piled up from self-service Freestyle soft drink fountains that allow customers to mix different drinks. There are more than 40,000 Freestyle units in the US, serving 14 million drinks in total every single day. Using the huge amounts of data it analyzed, Coca-Cola rolled out Cherry Sprite and Cherry Sprite Zero last year. And Greg Chambers, global director of digital innovation at Coca-Cola, says that in the future, AI will play an even more important part in the group’s vending machines and chatbots.
  • Harley-Davidson deployed an AI-powered marketing platform called Albert from the company Albert Technologies to pump up sales leads at a dealership in New York. They launched their first AI campaign, “48 Bikes in 48 Hours,” in early 2016 to reduce seasonal overstock and determine how quickly Albert could ramp up its audience knowledge to meet the brand’s conversion goals. Prior to Albert, the all-time sales record of the dealership was eight motorcycles in a single weekend. In its first two-day campaign, Albert almost doubled this, selling as many as 15 bikes.

AI will become a necessity for financial services if they want to deliver better experiences, lower costs, reduce risks and increase revenue, an executive at Infosys told The Digital Brand in its 2018 outlook. He also added that “most firms will need to evaluate a buy/partner decision to deploy solutions such as chatbots, biometrics, fraud and voice or else fall further behind.”

Another executive from Adversitement predicts that ML applications will continue to mature, with each vendor featuring a domain-specific solution. “Organizations will work towards fully integrated, end-to-end data management platforms to handle increases in different data streams, including deep learning applications, transforming data into actionable insights. AI and Deep Learning applications in voice recognition and video analytics will also accelerate,” he says.

Machine learning: banks go for big wins

Digital leaders in the banking scene have already made considerable success with AI and ML, as shown by industry examples gathered by Marketforce and the Bank Administration Institute (BAI), a US non-profit organization advising financial institutions:

  • Bank of America is planning to roll out Erica, its virtual assistant, to customers in 2018. Now being tested with employees, the chatbot will use predictive analytics and cognitive messaging to provide 24/7 financial guidance. Erica can also help with simple transactions, including checking account status or paying debt. Customers can also chat with Erica through voice or text message, similarly to Amazon’s Alexa.
  • JP Morgan has launched its Coin (contract intelligence) program to cut down on loan-servicing mistakes. This ML technology is used to review and interpret commercial loan agreements and can save an estimated 360,000 hours of human work normally done by lawyers and financial loan officers.
  • Wells Fargo piloted an AI-driven chatbot through Facebook Messenger in the spring of 2017. It’s aimed at delivering live information to help customers make better financial decisions. Using a conversational interface, the yet-to-be-named chatbot answers basic questions about account balances, payment due dates and customer routing number information.
  • The Royal Bank of Scotland (RBS) is using an automated lending process to approve commercial real estate loans up to $2.7 million in less than 45 minutes. This process would normally take days. The 2017 AI-driven launch is part of the bank’s broader digital and innovation agenda. RBS has also adopted a cognitive chat bot, powered by IBM Watson Conversation, to answer customer queries.

BAI predicts that 2018 could be the breakthrough year for AI. “It fuels creative thinking and action to drive cost efficiency with more intellectual horsepower. And it touches just about every area of the industry, from risk management, to the loan decision process, to talent management, to the creation of positive, powerful customer experiences,” BAI president and CEO Debbie Bianucci believes.

AI and ML in digital banking sales