How AI is Revolutionizing Banking: From Customer Service to Fraud Detection

Financial institutions are leveraging AI for improvements across multiple domains

Bankers are keen on AI's ability to bolster their operations in spheres ranging from customer service and regulatory compliance to research and enhanced security, even as they are mindful of ensuring AI does not inadvertently debase the customer experience, and that customers essentially can always reach a "live representative." Overall, the use of AI in the banking world is breathtaking, with 98% of North American institutions implementing AI for at least one operational process as of 2025, according to CoinLaw.

Customer Experience

In the customer experience arena, Valley Bank's Executive Vice President and Chief Operating Officer Russ Barrett broadly notes that customers have varying levels of "digital trust" and that his institution is committed to "never [compromising] the fabric of that trust." He adds, "[AI] is more like [additional] optionality for the early [bank customer] adopters, but we're never going to be able to force people to go at a pace they're not comfortable with."

At some banks, AI's assistance may be multi-layered: While AI can answer customers' routine questions and assist with tasks such as establishing accounts, bank employees can separately place customers' more challenging questions (received, say, over the phone or in-person) into AI systems that return sophisticated answers.

Within the context of financial institutions' evolving policies and internal systems, Barrett explains, "Rather than an employee [having] to constantly poke another [bank employee] or try to call someone to get the answer, [employees] now have an [AI] assistant that will be able to field a large number of those inquiries" without having to place customers on hold or call them back.

At certain banks, AI can also proactively reach out to customers. For example, whereas bank customers of yesteryear might use their last paper checks before considering ordering new ones, AI can now potentially understand that the last checks have been used and automatically contact customers regarding ordering new supplies.

Research and Preparation

AI can additionally help with larger business-client tasks. "AI is like having an assistant who spent the last two to three days researching your client before every meeting that you have, which is really neat," explains John Kowal, senior vice president, managing principal, and chief technology officer at Bedminster-based Peapack Private Bank & Trust. "That's the level of preparation we want for our bankers. I want our bankers to walk into their client meetings with an expert level of knowledge about the client, the prior meetings the client had [with the bank], the client's finances, and a plan for what's next. AI is helping us prepare that much faster than we ever could."

Security and Compliance

AI can also aid in a range of security scenarios. Against the backdrop of Federal Trade Commission (FTC) data revealing that consumers reported losing more than $12.5 billion to various types of fraud in 2024 (representing a 25% increase over the prior year), AI helps banks improve their ability to detect and respond to both fraud scenarios and money laundering.

"We're good at detecting illicit activity like money laundering, but what we're now starting to do is detect activity that might indicate our clients are becoming victims of targeted scams," explains Peapack's Kowal. "That's important to us because AI's ability to learn about tactics and detect them in real time [as well as] find indicators that our clients might become victims of these scams is really powerful ... it's going to make life a lot more difficult for those people out there who are looking to steal money from our clients."

In the compliance realm, manual reviews are becoming automated and thus being completed faster and more frequently. "Instead of looking at sample selections [when] auditing, we can look at everything now because AI doesn't have a limit on how much data it can process, and we can get a more thorough look," Peapack's Kowal explains. "The human reviewer is still going to be there, but if AI can give the human reviewer a second look, we won't miss anything."

AI Risks

Incorporating AI into banking systems doesn't come without risks, and topics including, but not limited to, AI prompt security and securing client data within AI systems are paramount as the technology evolves. From a cybersecurity standpoint, Valley's Barrett explains, "The only way to combat the risks of AI is to use very sophisticated AI that would be able to counter and highlight when there is an anomaly or a risk in what we are seeing, in certain inquiries, activities and consumptions; things that are very different paradigm shifts from where we have been traditionally in cyber risk."

There are other risks. Barrett says, "Banks cannot afford [AI] hallucinations. Banks cannot afford data quality issues. Banks cannot afford a misunderstanding. What that means is that the pace at which [AI] might move in your day-to-day consumption of information probably would be very concerning if that's the pace it moved your money. The underpinning of what your experience will be with these technologies [requires] a lot of hard work [including] getting the data structured, organized, controlled and of really high quality to make sure your experience is 100% reliable and explainable."

Separately, Peapack's Kowal stresses the need for AI confidentiality. "AI can be done securely," he says. "You do not want to put any sensitive data into Google's AI using a public account, but if a bank were to partner with Google, draw up an enterprise agreement and validate the security, [then AI] can safely be used with banking information."

Early Adopters

While AI is evolving at many financial institutions, Bank of America can be seen as an early adopter in part because in 2018 the bank launched Erica® – "the first widely adopted AI-driven virtual financial assistant" – within its mobile app, which has thus far been used more than 2.5 billion times.

Bank of America says AI and machine learning technology is embedded across the company's business lines in areas including global technology, operations, and training.

"AI is having a transformative effect on employee efficiency and operational excellence," Bank of America Chief Technology & Information Officer Aditya Bhasin said in an April 2025 statement. "Our use of AI at scale and around the world enables us to further enhance our capabilities, improve employee productivity and client service, and drive business growth."

The use of AI in the banking world continues to evolve overall, and also within the realms of predictive analytics and automation as well as investment and portfolio management. While AI's future industry-wide applications aren't entirely known, Peapack's Kowal concludes, "In the industry, we're known as a technology-forward bank, but our goal is not to be the most AI-driven bank out there.

It's to be the most trusted and responsive partner to our clients. I think AI helps us get there faster. So, we keep our eye on that end goal, and AI is helping us in ways I think we couldn't have imagined."