Machine Learning for Better CX

By Greg Ablett

Machine learning will make banks more personable.

Banking and financial services companies have made significant strides towards creating more personal user experiences. But in an industry where competition for customers’ attention is fierce—yet information security is of the utmost concern—evolving to keep up with expectations for choice, convenience and control requires thoughtful investments in how data is consumed and applied.

How are customer experiences growing more customizable?

Data has already helped transform how banks interact with their current and prospective customers. From the ability to use touch ID to sign into mobile banking apps to instant customer recognition, banks are advancing the ways they use data to enhance self-service—diminishing the need to visit a bank branch. Every transaction consumers make, each peer-to-peer payment solution we use, and the frequency of our ATM withdrawals provides banks more insight to predict our future behavior and tailor our experiences.

However, even with a wide variety of data points and metrics (gender, location, account history, time of day, payment type, etc.), most banks have funneled customers into a finite number of predetermined audience segments, based on assumptions about how they want to be serviced. Now the traditional relationship between banks and consumers is on the verge of a major shift: Business analytics and machine learning will help financial institutions capitalize on contextual intelligence to dynamically personalize interaction for each individual in real-time.

Machine learning: heralding a new era in customer experience lifecycle management.

Machine learning has great potential to improve the customer experience (CX) at today’s banks by taking advantage of the data already at their disposal. When deployed correctly, this technology can predict and actively address customer questions and concerns before they arise. As a result, customers can dodge common annoyances like navigating unwieldy banking websites or complex phone support menus.

Let’s look at interactive voice response (IVR) as an example.

Traditionally, a caller might hit a main menu and, based on the initial selection, be introduced to one of several common call-flows—arriving at resolution (hopefully) after fielding a slew menu options. But as banks introduce machine learning, they’ll examine historical data to create a library of individual call prompts that address common reasons why a customer would require service or support. Then, they will leverage a gamut of contextual insights around interaction type (appointment confirmation, bill pay, etc.), location, demographics, and past behaviors to automatically pinpoint the optimal sequence of prompts for a specific caller, at a specific moment in time. To make self-service quicker and even more personal, algorithms adapt to every customer and interaction while continuously learning from themselves to enhance future sequences.

It’s important to note that machine learning isn’t isolated to individual customer channels like voice, web, or SMS/text. Rather, it’s most effective when channels are well-integrated, and data and the customer experience persists among them. For example, once a customer pays his credit card via an IVR system, he may receive an automated text message to confirm the transaction. Likewise, a machine learning-enabled communication environment may note that an inbound call is from a customer who regularly makes an online account transfer at this time of the month, and may automatically route them to such a prompt that conveniently skips the other menu options.

Navigating new dimensions of data security and compliance.

As mobile technology continues to advance customers’ expectations of banks and other consumer brands, now is the time to leverage the power of machine learning—and carefully consider the other ways it will inevitably impact business.

Imagine this.

A customer accesses her smartphone with a PIN, and opens up a mobile banking app with a similar multi-factored login to check her account balance. She then logs into her account on her cable provider’s website, where she has a monthly payment due. Between these two active sessions, the customer has confirmed her identify several times. But what if her smartphone could save her time and verify this information for her, using her past behavior to intuitively prompt “pay July cable balance” based on a single thumb-impression?

When enabled by machine learning, the same customer’s credentials could be securely passed across all of her routine banking interactions—eliminating the need to repeatedly verify her identity while helping dodge risks like false identification, fraud and insufficient access.

Though these innovations simplify the banking experience, they also invite new security concerns with confidentiality, regulatory compliance and data privacy. It will be up to financial services organizations to establish and maintain robust data handling and protection programs that safeguard customers’ personally identifiable information. Organizations should replicate these security processes across all business divisions, and hold external partners and vendors accountable to the same standards.

Achieving success through machine learning.

To truly benefit from machine learning, banks need to paint a big picture, but start with small steps. Before envisioning its effects on CX, it is crucial to facilitate complete visibility and connectivity between all organizational functions, processes, communication channels and, of course, data points. Banking leaders should try machine learning on an audience subset or interaction type before rolling initiatives out across every audience and channel.

Part of the beauty of machine learning is that it’s built on the principle of continuous trial and error. Weaving machine learning into banking customer experience won’t happen in one sweeping implementation. It’s an ongoing project that can strengthen banks’ customer service and even overarching brand reputation over time. Banks are already math machines, so why not add some machine learning to their arsenals?

Greg Ablett is senior vice president at West Interactive Services, a provider of innovative customer experience and technology integration solutions. Greg oversees utility business and professional services, including business analytics, user-experience design and speech teams. Email: [email protected].

Online training in digital, mobile and social media from ABA.