Data insights are growing in importance for bankers seeking not only to better understand their customers, but also to explore new products and services — and many are using artificial intelligence to do so. But new research from American Banker finds that trust is still holding many back from embracing the technology on a broader scale.
American Banker surveyed 130 bank and credit union leaders for its
Top findings from the report
- Risk and compliance, financial analysis and human resources analytics were the
top three testing grounds for AI tools. - National banks saw
fraud-detection efforts as the prime area for AI to drastically improve over the next 12 months. - While national banks were the largest adopters of
chatbots , other respondents favored customer biometric-focused use cases. - Credit unions and regional banks were more likely to say that data systems
weren't precise within their organizations. - Most reported that their data systems were
adept at capturing useful data , but community bankers had the highest share of negative responses.
Where is AI heading to next?
Results from the research are highlighted below using interactive charts. Mouse over each section for more detail, and click on the chart labels to show or hide sections.
Bankers are deploying AI throughout their organizations, from risk and compliance to sustainability and recruitment efforts.
Risk and compliance were the top use cases: 59% of bankers and credit union leaders said they are integrating and testing new AI-powered tools for these purposes. Almost half (49%) said they were using AI for financial analysis and modeling.
Other top uses included human resources analytics and workforce planning (48%), supplier management (38%) and sourcing and contracts (38%).
Financial advice could take the top spot in the coming years however.
Jelena Zec, director of venture investing for Citi Ventures, said, "one of the most promising areas for AI adoption is wealth and retirement advisory services," as tools dedicated to reducing friction points when interacting with clients would have a significant impact on personalizing the whole experience.
"These tools are enabling a future where retirement portfolios can be continuously optimized using behavioral data and market signals," Zec said. "They're building the infrastructure for more intelligent, responsive financial planning across institutions."
Lending a hand
Financial institutions view AI as both a tool for fraudsters and a means of protecting against fraud.
National banks, community banks and credit unions all identified fraud detection as the top retail banking area where AI will have the most impact over the next 12 months, garnering 65%, 52% and 77% respectively. Fraud detection ranked third for midsize banks.
Ongoing risk management, regulatory compliance initiatives, financial advice for customers and customer service processes were all top retail banking activities identified by respondents as ripe for the addition of AI.
Experts with the global consulting firm Baringa say that while AI adoption and integration is in a growth stage across the banking sector, bankers' low appetite for risk is holding back full-scale adoption of the technology.
"Given the propensity for generative AI to hallucinate, 'narrow' AI techniques continue to be more frequently used for tasks that need to be deterministic and explainable, whereas generative AI is being used most commonly for tasks where a narrative is needed to explain things," said Brad O'Brien, a partner in Baringa's financial services practice.
With proper oversight, AI adoption is diving into areas like risk and compliance, which O'Brien identified as a burgeoning effort driven in part by both "reducing the administrative burden of capturing and recording qualitative risk and compliance information" and "enhancing predictive insights to help mitigate risk before it crystallizes into an incident or event," he said.
Tools of the trade
Most banks are keen to adopt AI in some form, but not all agree on what the top form to integrate is.
A majority (70%) of big-bank respondents said they have fully integrated AI-based chatbots. Biometrics such as facial and voice recognition were a close second at 63%.
Among the other three categories of respondents, biometrics were the top AI-based technology implemented across the organization, coming in at 59% for regional bankers, 32% for community bankers and 23% for credit union professionals.
For Katie Rice, senior executive partner of banking, financial services and insurance for consulting firm Ranstad Digital, instances such as Bank of America's "Erica" and Capital One's "Eno" are proof enough that chatbots and conversational AI have become table stakes.
"Organizations that are able to enhance their customer experience through these channels are providing instant support and more 24/7 availability without waiting for a human agent, improving customer satisfaction and at the same time reducing operational costs which continue to be [at the] top of big banks' minds," Rice said.
Where are banks' and credit unions' data systems falling short?
Data infrastructure is a crucial component of any financial institution, responsible for processing and recording transactions, managing account information, parsing custom insights and more. Some are better equipped than others, however.
Roughly 80% of national banking respondents said their data systems were accurate to some degree, against the 20% who said the opposite. Community bankers were just as optimistic about their data systems as their largest counterparts, recording a similar 80%-20% split. Regional bankers were marginally less optimistic about their data systems, with 36% saying their organization's systems were not precise, 63% saying they had some level of accuracy.
Credit unions were the group most evenly split, with 38% responding not precise against 61% who said accurate.
Data science experts say that AI tools can help close these gaps and improve the efficiency of in-house data systems for banks and credit unions alike.
Chester Leung, cofounder and head of AI platform at the cryptographic AI acceleration firm Opaque Systems, said that trust issues over the security of customer data handled by AI platforms hold back institutions from digging deeper with the technology.
"In the financial space, [this] concern is more than justified, [as] financial data is highly sensitive, and misuse or leakage has huge implications," Leung said. "But without AI that can aggregate, filter, generate and predict over large datasets faster than humans can, banks are leaving significant value on the table."
"The ability to surface insights at scale, including risk signals and consumer behavior patterns, is critical, but right now it's gated by trust," he said.
Data rich, insight poor
Alongside accuracy, the ability to capture and compile relevant customer data is a key factor determining how successful a data system is.
More than 70% of national bankers said their data systems were able to capture all relevant data from account holders and elsewhere. Credit unions were a close second at 69%, followed by regional bankers at 64% and community bankers at 60%.
Among community bankers, 40% said their data systems weren't capturing any relevant data. Among credit union respondents, none said their organization's data systems were unable to capture relevant data.
Erin Rahaim, AI compliance officer for the cloud-based digital banking provider Alkami, said banks and credit unions alike that are "data rich but insight poor" stand to benefit from using AI to better understand troves of consumer data and provide nuanced insights to their customers.
"FIs that can leverage AI's predictive capabilities to anticipate account holder needs, such as reminding users about savings transfers, are ahead of the curve in engaging with their account holders in a more meaningful way," Rahaim said.
