The Two Biggest Roadblocks to Bank Automation

William Heitman, Partner & Managing Director, The Lab Consulting

William Heitman, Partner & Managing Director, The Lab Consulting

The future of banking as imagined in the professional journals features digitization, robotics and artificial intelligence leading to lavish improvement in productivity and customer engagement. Turn the page, of course, and there you read that the future is actually digital financial startups, or fintechs, bringing the imminent demise of banking as we know it.

Relax. In 1994, Bill Gates called banks “dinosaurs” and predicted that technology would drive them to extinction. Today banks are still operating pretty much unchanged. The technology has certainly changed, but the way employees work hasn’t, so there has been little improvement to cycle times, error rates, and processing costs. A homebuyer, for example, will wait about 50 days to close a mortgage that requires only 12 hours of “touch time.” All of the new technology has changed nothing.

In the past two decades, by comparison, automakers reduced their assembly labor by roughly half, while increasing the useful life of cars by more than 35 percent. No one has similarly discovered how to improve productivity or reduce operating expenses in banks.

There’s a reason for this. Executives who manage white-collar, or knowledge, workers don’t see similarities with factory work. That’s unfortunate because non-technology industrialization—standardization, simplification and the division of labor—holds the key to automation.

Management had an early chance to industrialize the office. Large-scale office automation technology first appeared in the 1920s: mechanical filing systems, calculators, punch-card tabulators. These could have delivered massive cost reduction in banking. Vendors promised that this “early IT” would deliver breakthrough productivity gains and cut costs. Management visionaries like W.H. Leffingwell and Lee Galloway demonstrated that the same industrialization approach used in the plant would work in the office.

But the promised benefits required that office work first be standardized—industrialized—and executives resisted.

Back then it didn’t matter much. Knowledge workers represented a miniscule portion of company value. Today the activities of knowledge workers represent an average of 60 percent of the value of the typical S&P company. Today it is critical that banking executives ignore convention, make the perceptual leap and accept that industrialization is not just for automating factories.

You can begin by navigating around two, highly effective automation roadblocks.

Automation Roadblock I: Standardize Ambiguous Operations

An internal team documenting a mortgage-servicing department’s business processes identified five end-to-end processes that covered over 85 percent of all the work activity. The sales group pointed out, however, that a different team had identified more than 1,400 processes. Neither team’s definition was right or wrong. Ambiguity was the enemy of both.

This is not unusual. Knowledge work organizations typically have no procedures to standardize the definition of their operations. That means that people will often talk past each other, thinking that everyone is on the same page. In practice, they are miles apart.

The unstandardized definition of something as seemingly obvious as a business process is often wildly inconsistent—within the same company.

For example, in a global securities brokerage the IT group proudly announced their heavy investment in business process automation. However, they categorized “currency conversion” and “invoice generation” as examples of the type of business processes they were currently automating. Dramatic gains in productivity for these “business processes” were promised and delivered. They were disappointed that their successful efforts seemed to be barely noticed by the business line executives.

That’s because the business line executives view these as transactions, very small parts of a more comprehensive “business process,” such as customer product on-boarding, order-to-cash or record-to-report. Line executives wouldn’t disparage IT efforts to automate transactions. But they won’t be holding their breath for the type of productivity gains that make a competitive difference to their business. Reducing ambiguity between IT investment plan and the needs of the business executives could have avoided this costly, “ho-hum” outcome.

How do you avoid this roadblock? Avoid assumptions. Define everything, especially if it seems embarrassingly obvious. Steal the techniques used by any factory operation. Use diagrams where helpful, just as they do:

1. Create a standardized dictionary of terms common to the various groups involved.
2. Store this dictionary in an easily accessible repository and publicize the location.
3. Define a process to manage changes and additions to these standards.

Automation Roadblock II: Mandate Design for Machinability (DFM)

After the introduction of assembly-line robots in the early 1980s, it only took about a decade for automakers to produce selected components in fully automated plants. How? They mandated that engineers redesign products to make the work compatible with the capabilities of the new machines. Then they piloted and continuously refined the products, the work, and the machines. After a few initial missteps, errors plummeted while productivity, quality and reliability skyrocketed.

Contrast that to the implementation of workflow automation and enterprise software in the office. It’s been three decades since their arrival, and yet productivity gains and cost reduction have been elusive. That’s because office work and its products remain relatively un-designed. Knowledge work businesses like banks search for machines—technology—that are compatible with the existing work. Use of this technology remains optional. Employees use it when compatible and convenient. Where the work and machines are incompatible, they await the next generation of new technology. 

A regional bank, for example, installed new workflow technology with an automated underwriting module for its consumer and small business loans. Besides improved productivity, a major goal of the new system was more rigorous compliance with federal anti-money laundering (AML) guidelines. The problems began with AML. The bank’s compliance organization could never seem to collaborate with the operations group to agree on and document the “work products” required to meet the AML specs. Consequently, the IT group had a difficult time figuring out which features and functions to purchase in the new technology. Deadlines loomed and everyone simply pressed on.

The result was a massive investment in high-powered, under-used technology. Workers were not mandated to use the workflow tool. Consequently, they circumvented it with emails, texts and phone conversations. Although most of the work ultimately ended up in the system, the start-to-finish audit trail of documentation—for compliance as well as productivity monitoring—was woefully incomplete. Lending officers, and ultimately customers, were supposed to be able to access the status of pending loans via mobile devices. With so many loans and progress updates entering the system late in the process, however, this feature was unreliable and generated little confidence among lending officers. So they simply picked up the phone and called underwriting, just as bankers have done for a century.

To dodge this roadblock, mandate that knowledge work be designed to be performed by machines (not people). Extend the standardization work from Automation Roadblock I:

1. Invest to create an “industrial engineering” team to study every aspect of the “knowledge work factory” that is your bank. Unprecedented? No, this is how Ford executives created their moving assembly line.

2. Define finished “work products” and work backward to engineer “bills of material” and production instructions for each.

3. Work hand-in-glove with the IT team to refine the design of these work products so that machines can perform as many activities as possible. Test, refine and measure productivity gains. Repeat forever.

Banks are easier to automate than auto plants. Your P&L, and your customers, deserve no less.


Weekly Brief

Read Also

 Bank Security through the Years

Bank Security through the Years

John Deerin, SVP, Security Director & BSA Officer, The Bank of Tampa
Industry-focused Business Intelligence as a Service

Industry-focused Business Intelligence as a Service

Frank Tian, VP, Risk Management, MUFG Union Bank
Data Analysis in Storytelling: Common Mistakes

Data Analysis in Storytelling: Common Mistakes

Albert Chin, Head of Model Risk Management, Signature Bank
 Revamping Regulatory Practices with Innovation

Revamping Regulatory Practices with Innovation

David Cowland, Head of Compliance Operations, Fidelity International
Why we need MSSP? Managed Security Service Provider & the quest for cybersecurity experts

Why we need MSSP? Managed Security Service Provider & the quest for cybersecurity experts

MarekTrebicki, Cyber Security Services Risk & Control Associate Director, Standard Chartered Bank