Bank and credit union leaders are eyeing intelligent automation as a path for growth and improvements in fraud detection, customer service and other areas, according to new research from American Banker. But these pathways aren’t without their challenges.
Stifling company cultures and data security concerns are among the obstacles hamstringing the pace of automation adoption across the financial services industry.
American Banker surveyed 153 bank and credit union leaders for its
Top findings from the report
- The
C-suite tends to have more decisioning power than boards of directors in matters of deciding automation plans. - Banks and credit unions are either
staying the course or doubling down on investments in intelligent automation. - Data privacy and incompatibility with existing IT systems are
major hurdles to furthering adoption of tools for automation. - Additional help is
often needed to support automation efforts, including training and education. Fraud detection is a dominant use case for automation across the next three to five years.
Results from the report are highlighted below using interactive charts. Mouse over each section for more detail, and click on the chart labels to show or hide sections.
When it comes to who has the most control over choosing an organization’s intelligent automation pathway, power sits at the top.
More than 60% of respondents said that business-focused C-suite leaders including chief executives and chief financial officers were the final deciders in their organization’s intelligent automation decisioning process. Among only 33% of respondents, tech-focused C-suite leaders such as chief information and chief security officers were the deciders.
Involvement seemed to be more evenly distributed as you stray further from the C-suite, from recommendation and input providers up to those who are simply informed about decisions being made.
For efforts that span across multiple departments, determining not just who has oversight of the technology but maintaining a holistic view of progress can be tricky.
“Developing and keeping an end-to-end view of a process is critical, and often difficult for FIs with complicated organizational structures and overlapping interactions of different areas with a given workflow,” said Christopher Miller, lead analyst of Javelin Strategy & Research’s emerging payments practice.
“The impetus for solving some problem comes from an area that is within a larger workflow and this team might not be best situated to choose a solution that enables the real gains to be realized,” he added.
Starting an intelligent automation project requires “clear goals” at the start, according to Patty Corkery, president and chief executive of CUSG and Michigan Credit Union League & Affiliates, who added that other priorities should be “strong oversight in place, good documentation practices and [assuring that] any new technology that you introduce fits with your existing systems.”
Banks and credit unions of all asset sizes are either staying the course or doubling down on their tech spending for automation tools over the next 12 months.
Regional bankers, about 36% of respondents surveyed, said their organizations were planning to significantly increase spending on automation in the coming 12 months. Twenty-six percent of national bankers said the same, as did 15% of community bankers and 5% of credit union professionals.
No respondents said their organizations were planning to cull tech spend to any degree.
While spending is by and large on the rise, consultants in the financial sector say that banks and credit unions alike can start to best deploy automation and AI tools in two areas: daily workflows and data-intensive processes.
“Embedding AI into core business functions, upskilling staff to work alongside these technologies and establishing strong governance frameworks are key to unlocking the full value of banks’ data assets,” said Chris Long, leader in KPMG US’s financial services consulting practice.
For Tom Ellis, head of mobile and associate facing channels technology at Bank of America, this mindset has remained a constant across all of the bank’s business lines.
“The same AI foundation behind our virtual financial assistant, which has supported 48 million clients through nearly 3 billion interactions, also powers internal tools like Erica for Employees, now used by more than 90% of staff to streamline support and cut IT desk calls in half,” Ellis said. “When you treat it as enterprise infrastructure rather than a collection of tools, the impact is both broad and lasting.”
Not all banks and credit unions are facing the same struggles when it comes to adoption.
Among large banks, 50% of respondents said competing internal priorities is a top hurdle to adopting intelligent automation; 50% also chose compatibility with legacy IT systems. For regional bankers and credit union executives, data security and privacy concerns (50% and 56% respectively) were the primary impediments to intelligent automation adoption.
Among community banks, 46% said resistance and company culture were the main challenges to intelligent automation for community bankers.
Michelle Boston, chief information officer and head of data management technology for Bank of America, said “prioritiz[ing] building a clean, connected data layer” at the start of an automation journey can establish a solid foundation for navigating and executing the implementation phase.
“Get your data house in order [first] before you over-rotate on AI,” Boston said. “When implemented correctly, intelligent automation becomes the infrastructure layer that powers an institution’s entire operating model. Intelligent automation is the engine that turns data into actionable insights, applies those insights at scale and continuously learns from every interaction to compound value over time.”
Data is a key determinant of how effective automation efforts are within a financial institution.
Pete Chapman, chief technology officer for Grasshopper Bank, said areas like lending especially “stand to benefit greatly from data structure and aggregation” that can cut down lengthy processes to minutes as opposed to hours.
“At Grasshopper, we’re using AI to automate annual reviews of our commercial loan portfolio, saving time and improving consistency by pulling relevant data into a structured format.” Chapman said.
Adopting automation tools is only the first step for many of these organizations, as many begin to shift towards building support networks for the technology and its users.
Roughly 50% of national bankers said when it comes to automation skillsets, help is needed in both process analysis and design as well as data preparation and cleansing.
For regional and community bankers, 56% and 57% respectively said they need more support in developing and training AI models before implementation. Fifty-two percent of credit union respondents also identified AI training as the top area in need of support.
Ami Iceman-Haueter, Chief Experience Officer for Michigan State University Federal Credit Union in East Lansing, said selecting use cases where “automation can make a measured impact” early on can boost buy-in from employee company-wide.
“Pick one or two areas where automation can make a measured impact,” Iceman-Haueter said. “That helps build confidence in the process and shows teams the benefits in their work. It’s also important to set up governance, training and support early on, so you have a strong foundation as you expand to more use cases.”
At Digital Federal Credit Union, a $12.7 billion-asset institution in Marlborough, Massachusetts, executives agreed that setting KPIs from the beginning can help show the value for those hesitant to wade deeper into automation.
“It can be difficult, especially in a field where crisp math is preferred, but setting KPIs from the beginning can help show the value,” said John Mason, chief technology officer for DCU. “At DCU, we measured the hours of labor saved by implementing intelligent automation and found it was 400,000 total hours over the past 18 months.”
Fraud is on the rise across the banking system, and executives are predicting that intelligent automation can help fight this rising threat over the next few years.
Across all respondent classes, fraud detection was the number one use case for intelligent automation that has the most potential to change how a bank or credit union operates within the next three to five years.
Other top applications range from regulatory and compliance tasks to customer service and support.
Stewart Watterson, strategic analyst for advisory firm Datos Insights, said that when speaking with retail bankers on what areas in need are getting the largest influx of investment, fraud and risk mitigation tend to be common areas mentioned.
“When you ask bankers either ‘where aren’t you currently using AI’ or ‘where are you working towards using AI,’ it’s generally fraud and risk mitigation that is named,” Watterson said.
The scale and number of automation products depends largely on the size of the organization as well as the sophistication of technology in use.
Benjamin Seesel, vice president of advisory in research firm Gartner’s financial services practice, said smaller institutions can be “AI- and tech-forward,” but with a comparatively limited budget against larger players, “clear criteria and decision-making tools” are a must for picking the technology that makes the most sense.
Intelligent automation inherently means handing over control of business functions to AI models in some capacity, but experts say be it accountability or explainability, human oversight is still essential.
When pursuing any use case in finance, it’s important to understand that many applications are less a “full handover of decision-making” and more “a partnership between humans and AI,” said Zoey Jiang, assistant professor of business technologies at Carnegie Mellon University’s Tepper School of Business.
“Workflows should be designed so that humans remain in the loop for complex or high-impact cases, with clearly defined roles, structured processes and override controls,” she added.