Enabling Technologies

The Role of AI Technologies in Humanizing Digital Banking

Envestnet | YodleeVice President of Product Applications

Digital banking has not only gone mainstream but in many cases, it has become the only point of contact between banks and their customers. According to a May 2016 report on the tablet’s shifting role in mobile banking, half of the US adult population now banks using smartphones and tablets (an increase of 29 million people from the year before). While financial services delivered through digital channels have brought the convenience of anywhere, anytime banking to consumers, what is the cost? And, are banks living up to the expectations of their customers?

One of the main challenges the financial industry faces is the risk of diminished customer relationships as we make the shift to digital. Let’s face it, online banking sites can be pretty impersonal. For many years, face-to-face interactions were the norm as customers would develop a relationship with a bank associate who knew about their financial situation and would help them reach their financial goals. The traditional bank branch is now used by just 32% of customers, meaning the opportunities to create personal relationships with customers has greatly diminished.

As an industry, this is where the application of artificial intelligence (AI) and machine learning (ML) can help turn the tide. The promise of AI-powered engagement tools is to provide customers with applications that understand what they want and actually help them do something about it. Imagine if your bank’s passive, impersonal digital banking portal turned into a virtual financial assistant that is singularly focused on helping you make the best of your financial health. An assistant that never sleeps; constantly running in the background, monitoring your overall financial data in real time and pushing out actionable insights about your finances and opportunities to take smart action. It helps you take steps to improve your financial health by recommending easy ways to save more, spend smartly and make plans for your financial future. Your AI virtual assistant would go everywhere with you, and you could interact with it on your cell phone by chat or through your connected home devices by voice. Finally, your assistant would not only monitor your historical data, but it could help to predict upcoming financial events and get you organized so you can deal with your finances proactively.

Does that sound like the stuff of sci-fi? Well, it’s not. The technology is improving by leaps and bounds. There are a number of companies that are working on exactly that type of experience. We may still be a year or two off, but when we get there, these technological advancements will lead to a democratization of financial health that can put banks squarely in the role of customer advocates. That’s the power of AI when done right. In the very near future, banks will be able to provide contextually relevant insights including peer benchmarks from similar customer segments, personalized specifically for each individual. They’ll be able to mine their network data for good financial behavior and predictors of success that can be shared with individual customers. Banks can become anytime, anywhere trusted advisors and advocates for their customers.

Smart financial institutions will also learn from the interactions that customers have with their AI virtual assistants. A big part of financial health is ensuring that customers take advantage of financial products that are optimized to their specific needs. Mining the interactions that customers have with their financial assistants can help banks offer the right type of financial products and services to meet customer needs.

Over time, experiences like chat or voice-enabled virtual assistants can provide a humanizing touch to the banking experience. Now is a good time to start experimenting with the technology and figuring out how you want to deploy it. As with all new technology, it’s best to start out slow and grow over time. Here are four considerations to keep in mind as you develop your initial pilots:

Start with specific use cases and remember to set expectations up front

When it comes to applications like voice assistants, it’s essential to set expectations upfront, so the user is aware of the parameters for what can and cannot be addressed. In early phases, voice commands were mostly geared towards very specific use cases.

For example, Apple faced a structural problem in its early days with Siri – no matter how well the voice recognition element worked, there were still only 20 things that users could actually ask. Apple gave consumers the impression that they could ask anything and everything and in turn, users often got a computerized shrug. Flash forward a few years and Amazon's Alexa tells a different story. The company proactively communicated what the voice assistant can and cannot answer. It also had the benefit of a much larger data set of behaviors to index from and react to. In essence, the system started off “smarter.”

Financial service providers need to start with some common, easy-to-respond to use cases. Clearly-defined, narrow domains like inquiring about bank account balances or transaction details also seem to work well, because users are able to easily identify what they can ask. And when you’re picking a provider, make sure the service allows you to seamlessly switch between a bot and a human assistant. Many of the better services have been trained to recognize customer frustration and transfer the interaction to a live representative.

Incorporate visual aids

Visual aids, such as a screen on a device can supplement virtual assistants and chatbots, cutting down user confusion by showing customers how the AI components of services work. Platforms such as Amazon Machine Learning offer visual aids and easy-to-access analytics to help businesses build machine learning models without having to create the code themselves. Other design elements to consider incorporating alongside AI-powered features include touch screens, icons and a selection of pre-selected next steps that allow the customer to easily navigate through the choices they can make.

Make sure your bot has a personality and soul

Most designers agree that bots need to have something that resembles a human personality. When done right, there's a mix of technical and creative designers working together to design dialogue – even when there are deeply complicated queries to process. For example, Google is hiring creatives to bring humor and storytelling to human-to-machine interactions, and Microsoft Cortana’s writing team includes a poet, a novelist, a playwright, and a former TV writer.

Chatbots work best when people interact with them like they would a human, so the machine can apply the conventions of conversation and not confront the user with unnatural syntax. Ensuring your bot has a personality encourages a natural back-and-forth dialogue.

Integrate with leading tech platforms

First movers in the AI space are likely going to be tech companies. Keep an eye out on what they’re doing and consider selecting platforms like Facebook Messenger and Alexa for your first forays. Customers can self-select into the service and you can work out the kinks before fully integrating services into your digital channels.

AI and ML have the potential to revolutionize all areas of our lives. For the financial services world, in particular, it has the power to help foster greater trust from consumers, which will be critical as more and more users shift to mostly using mobile banking platforms. As we explore all the possibilities these technologies present, it will be critical to place the consumer at the forefront. It is very likely that in the future, AI and ML will be commonplace in our day-to-day lives. Laying the groundwork now to ensure users are always the priority will undoubtedly help as the technologies start to become commonplace.

Katy Gibson

Envestnet | YodleeVice President of Product Applications

Katy Gibson is the vice president of product applications at Envestnet Yodlee. She has more than 20 years of experience in developing technologies that support financial institutions, focused on products and services that help users manage their finances.

Apply to Become a Contributor