What You See Is Only Half of What You Get With Digital Banking

As technology has become more personalized and ubiquitous, we’ve come to expect impeccable and instant digital experiences. To satisfy these customer demands, banks and other financial institutions (FIs) — like most enterprises — are rolling out new products and channels. This means cleaner mobile apps, more interactive chatbots and easier-to-use web portals.

According to research results from Cornerstone Advisors, 24% of the surveyed executives of U.S.-based, mid-sized FIs say their companies implemented new or updated mobile applications between 2018 and 2020, and 17% did so in 2021. Also, over half of them have at least discussed chatbots at the executive level, if they’re not already deployed.

Such user tools are now required to stay competitive. But it’s only half the answer. The customer experience (CX) isn’t derived solely from what’s on the screen but also how quickly actions from our clicks are resolved. Unfortunately for us as banking customers, although the front end is becoming more digital and efficient, the back end has remained highly manual. Analysts and operators fill their days combing through documents and navigating legacy applications. We may perceive it as a digital experience, but we’re often merely experiencing a digital façade disguising slow, costly, error-prone processes.

Why Much Of Banking Has Remained Manual

The customer lifecycle management (CLM) processes that banks use to support customers have remained manual because automating them is a difficult tech challenge. Within this set of processes are often outdated legacy apps and (maybe more importantly) a high variety of data.

Let’s consider opening an account for a business customer. To simply open a checking account, every limited liability company (LLC) needs to provide certificates of formation, personal identification for each member and additional details. Add more complexity to the business (such as multiple authorized signers or activities in high-risk countries) and the documentation requirements increase to include perhaps annual reports and beneficial ownership forms. Each piece of documentation has different data elements arranged in varying layouts, which makes this problem nearly impossible for traditional software to solve. Thus, the approach has remained unautomated, leveraging large operational teams to execute countless manual keystrokes and mouse clicks.

Better mobile apps, chatbots and online portals don’t solve this problem either. Yes, they simplify form submissions and document uploads. But when the customer hits that “submit” button, the screen reads something like, “Your submission is being processed,” not “Your account is activated.” Back-end analysts must still complete the workflow manually. And banking customers wait and wait.

Attempts have been made to increase operational efficiency. But those attempts often involve throwing more people on the team or deploying piecemeal IT solutions with minimal effectiveness (and they add even more operational and technical debt to an already complex operation and portfolio of systems).

Repercussions Of Slow Service

The implications of slow service can be hefty. To start, 10% to 15% of corporate banks’ annual revenue is lost to attrition. Customer dissatisfaction often starts during their first two weeks with a bank, which is, not coincidentally, the average time it takes to onboard a corporate customer. It’s not only slow but costly — each bank spends an average of $54 million for customer onboarding per year. 

Across the banking industry, customers experience slower service than they expect. Maybe it’s a real estate developer who needs a new account for every project and slowness adds time and stress. Or the tech startup that’s looking to scale quickly and doesn’t want to be distracted by yet another email asking for still more documentation. Or the fledgling, first-time business owner who may wonder whether life would be better at a different bank. This pain isn’t caused by the software itself but by the manual work initiated with its use.

Getting Beyond The Digital Façade

Machine learning offers a new operating reality for banks and other financial services companies. A breakthrough approach — embedded within automation and other software — enables a true digital experience for banking customers. When software can learn from data, platforms are able to manage the complex data that’s caused the challenges. The work that’s tasked to the organization remains the same, such as combing through and validating account opening documentation. But how it’s executed can become much more efficient because back-office teams are now augmented with smart software.

When starting with machine learning, it’s important to ensure the technical performance of its models. But relying on machine learning for everyday operations requires beginning with the end in mind to set machine learning in motion and answering the following questions.

1. What are the quality expectations from machine learning? How does this compare to operations today?
2. What’s the desired level of impact to cost, quality, speed and volume?
3. How will machine learning be operationalized for the first model and for all those to follow?
4. What controls will be in place to ensure performance, detecting any drift from expectations?
5. How will employees of the organization interact with machine learning output?
6. How will new data be incorporated to continuously improve machine learning?
7. What accelerators (e.g., prelabeled data and prebuilt solutions) are available to expedite implementation times?

Understanding these areas upfront provides an excellent foundation for readying the organization to work with machine learning, ensuring a more positive impact on the organization and its customers.

Today’s banking customers demand optimized experiences. Yes, we want better ways to interact via mobile, chat and web. But we’re unsatisfied by just a slick UI: We want the instantaneous resolution of why we’re engaging with the bank in the first place. The days and weeks traditionally required to execute customer lifecycle management areas such as account opening are far from instant. Machine learning-powered automation can help address the issues around data and systems in these processes, allowing FIs to turn their digital façades into real digital experiences.

This article originally appeared on Forbes.com.