For the past three years, Wells Fargo’s automation strategy has pushed the $1.96 trillion bank to invest more heavily in bots.

“Human doing it today, bot doing it tomorrow,” is how Paul Krauss described the bank’s early thinking on automation. Krauss is the enterprise robotics and intelligent automation leader at the San Francisco-based bank, where he heads up its robotic process automation Center of Excellence (CoE). That strategy is set to change in the coming years, as Wells Fargo hones its playbook to boost return on investment.

Bending the maturity curve

Since its inception in 2018, Wells Fargo’s automation team has grown to nearly 120 in staff, Krauss said. He likens the operation to a brain: The “left side” of the team leads strategy, technology roadmaps, lifecycle and methodology, while the “right side” focuses on developing the code, design deliverables and technical support. Members of the engineering and software side outnumber those on th strategy and business side about two-to-one, and the numbers are set to grow in the coming year, if job postings are any indication.

Wells Fargo has posted 16 automation-related job ads on its website since yearend, and the roles range from automation engineer to consultants for a U.S. regulatory reporting automation team. The bank is also working through a roadmap of more than 250 efficiency initiatives slated for the next four years, expected to amount to more than $8 billion in potential gross savings, CEO Charlie Scharf noted on the bank’s Jan. 15 fourth-quarter earnings call.

Engineering power aside, Krauss said it’s crucial for the CoE to tie technology ventures to business impact. That practice has prompted the bank to rethink its automation strategy as it looks to accelerate and scale its ROI by placing more emphasis on attended bots than on unattended bots.

Attended bots, also called “helper bots,” are those used by humans, like the chatbots used by call center employees to better assist customers by phone. They optimize tasks by offloading portions of each to complete the work more quickly than humans. Unattended bots execute tasks and interact with applications independent of human involvement. They can be triggered by events or scheduled to perform tasks, such as receiving a batch of customer information and automatically entering data points into various applications.

Wells Fargo, like many other big banks in 2018, started its automation program with unattended bots, which were relatively cheap to make. On average, a bot could be developed in eight weeks with $50,000, and the bank viewed it as a permanent self-sufficient fixture, rather than a technology that would require human intervention, Krauss said. Perhaps counterintuitively, that view shifted when the bank considered automation at scale.

“Everybody had this key change, probably about a year and a half ago, to say, ‘Okay I’ve got thousands of contact center representatives; if I could put bots in place that could simplify things for them — signing out, transferring calls, things that shorten average handle time — now I can impact tens of thousands of people at a time, rather than one,” Krauss said. “That’s where you started to see the shift toward attended, and why we probably still have more automations that are unattended,” he said, adding that investment is now shifting to attended automations.

Giving customer-facing employees automation tools, like a newly introduced attended bot, could save time for thousands of employees. Each employee delivers the time-savings to customers, thereby improving the bank’s customer experience. Ultimately, the ripple effect of one attended bot extends farther than the potential of one unattended bot.

Wells Fargo has thousands of attended bots and hundreds of unattended bots that save the bank millions of hours on an annual basis, Krauss said.

Corporate banking revamp

Wells Fargo is now focusing its automation efforts toward onboarding within commercial banking and treasury management.

Currently, Wells Fargo is piloting a new onboarding workflow within commercial banking that is designed to reduce the number of employee handoff and interactions with customers, said Allysun Fleming, senior vice president of treasury management and payment solutions. The bank is currently collecting customer feedback that will inform the bank which systems need to be integrated and rewired.

Before the bank can deploy any bots, it must rethink the process it is looking to automate, Fleming said. The challenge is to consolidate the number of employees required to interact with a customer during the onboarding process to just one person; currently several groups of employees are required for the task. Wells Fargo is in the design stage of mapping out the new roles and responsibilities, as well as the technical integration required on the back end, she noted.

Meanwhile, the bank is also exploring new data techniques that will allow customers to self-service while onboarding new or different products within the bank’s treasury management division. This will require the bank to simplify the data and translate the internal bank vernacular into terms the customer can understand and put into context. While the bank continues to iterate its data management and data terminology, the end goal would be for customers to duplicate existing offerings to new accounts after acquiring new assets, with automation speeding the delivery of the request to complete overnight, Fleming said.

The future of Wells Fargo’s strategy

If unattended and attended bots were the first two steps along Wells Fargo’s maturity curve, intelligent document capture and the use of cognitive technologies, like machine learning and natural language processing, are the third and fourth steps.

The more attended and unattended automations the bank introduces, the more reliant on data it becomes. As such, Wells Fargo is looking to beef up its automation capabilities around intelligent data capture, said Vibha Patil, the bank’s enterprise operations transformation leader. Patil works with the CoE, among other bank groups, specifically around process improvement and process modelling.

Specifically, the bank is looking to integrate into its tech stack more complex technologies, like motion capture engines and optical character recognition tools, to pull information off documents and input that data into systems, Krauss said.

The fourth step on the automation maturity curve overlays cognitive capabilities onto automated processes to mitigate the human intervention found within end-to-end processes. Wells Fargo is exploring various use cases of machine learning and natural language processing, but the development is nascent, Krauss said. The technologies require skill models, which in turn require a lot of data, he added. Wells Fargo will need a combination of those technologies, however, to reach a fully automation-enabled process, he added.

Looking forward, Wells Fargo will continue to advance along all four paths of automation development, Patil said. In addition to building automation capabilities in-house, the bank will also explore emerging fintech tools that facilitate processes for dealing with large amounts of data, such as reconciliation. Another future focus area will be the use of tools to connect some of the bank’s legacy systems that are not yet linked, allowing the bank to integrate its automations and unlock additional efficiencies, she said.

“Everyone today talks about digitization, but where I’m seeing the magic happen right now — it is in digital — but it’s really the combination of digital plus data management plus automation,” Fleming said. “All three have to come together to deliver something special for the customer, especially as our expectations continue to increase.”

Bank Automation Ignite, on April 13-14, is the event for inspiring automation initiatives and investment in financial services. At the virtual event, financial services professionals can discover new use cases and technologies that are accelerating automation in banking. Learn more and register at www.BankAutomationIgnite.com.