How AI is transforming the future of banking

February 04, 2025

Imagine this scenario: You glance at your watch and see a reminder about your daughter’s birthday next week. Based on the intelligence your bank has about you, your family members and the shopping preferences of girls your daughter’s age, a variety of appropriate gift ideas are generated. You choose one and the purchase is executed against your bank-issued credit card and delivered a few days later.

As we explore the future of artificial intelligence (AI) with our clients, we find a mixed bag of reactions to the possibilities ahead. Some can’t wait to be on the forefront, blazing trails with innovations their customers can touch and interact with. Within the next 12 months, some of our clients are on track to deploy customer-facing AI solutions designed to unlock how money is stored, moved and put to work. They know full well that there will be missteps and much to learn, yet they’re willing to assume the risk and push the envelope.

The cautious majority

Working with the early adopters among our client base is exhilarating, but they’re in the minority. More often, our banking clients react with caution, concerned that customer-facing AI could lead an accountholder down the wrong road or violate regulations around consumer protection, truth in lending, tax advice or other compliance-related blunders. In an environment accustomed to standardized procedures and carefully crafted scripts for every conceivable customer situation, these bankers have justifiable questions about how to monitor AI and control its behavior.

Make no mistake: This group is intrigued by AI and eager to embrace the promise of better operational efficiencies, fraud prevention and compliance. On the operations side, clients are asking how AI can help them become faster and more precise at responding to customers while lowering costs. On the technical side, we field questions around utilizing AI to generate, test and debug computer code faster, eliminating the manual element and enabling developers to function at a higher level.

They’re excited, but they don’t want AI to touch their customer, yet.

An air of optimism

Still, everyone is talking about AI and optimistic about the possibilities of both applied AI (used for problem-solving and performing tasks) and generative AI (used for creating original content).

According to the 2024 FIS® Global Innovation Research¹, more than three-quarters of financial services leaders expressed optimism. And they intend to take action.

In the U.S., 72% of financial services executives said they plan to increase their spending in applied AI.¹

In the U.K., 59% of financial services executives said they plan to increase their spending in applied AI.¹

Similarly, 68% of financial services executives in the U.S. plan to increase their spending in generative AI, compared to 57% in the U.K.¹

As for today, nearly half of respondents from the financial services sector report that they’re already deploying AI, with another one-third planning to use it within the next year. When asked about their usage, more than two-thirds of C-level leaders said they are utilizing AI and machine learning to reduce operational risks, specifically in the areas of security threat detection, risk modeling and predictive analytics. But their usage doesn’t stop with risk concerns. Roughly half are employing AI to understand customer preferences, predict market trends and better tailor the customer experience.¹

Data – The fuel that powers AI

As humans go through life, we’re bombarded with thousands of life experiences – or data – that help us learn about the world. In the same way, huge volumes of data are the fuel that AI relies on to learn, adapt and make informed decisions. But not just any data: It must be high-quality data with no built-in inaccuracies or biases that could lead to unreliable AI outputs. This mountain of structured and unstructured data often must be cleaned, formatted and prepared before being used to train your AI models.

Historically, data has not always been presented in a formulaic, structured way, which has made it difficult for banks to digest and use it. But large language models that have been trained on massive amounts of data to learn how language works have made interrogating unstructured data much more accurate. This innovation not only aids in the cleansing of data, but also enables banks to develop chatbots that go beyond basic questions and answers, making them more conversational, informational and even able to analyze sentiment. While useful in the customer service arena, this technology is also invaluable for sharing information internally throughout the institution.

Given these challenges and breakthroughs, you must be rigorous in surfacing and sourcing data, assimilating it and utilizing it to arrive at meaningful insights that inform better business decisions.

Move past the days of measuring the health of your organization merely in terms of sales and cost dynamics. Use your data to gain clarity by drilling down to the finer indicators of how quickly you fulfill customer requests, pay your suppliers, settle disputes or handle credit issues.

By taking an arduous approach to data analytics, you gain a 360-degree view of your entire organization, even if you are in multiple geographic locations, house a variety of core systems and operate a complex web of business units. On the customer service side, optimizing the use of your data is essential for creating positive experiences, solving pain points and forging deep bonds.

Explore the possibilities

AI is top of mind, and our banking clients want to be in on the action, or at least not left behind. But AI just for the sake of AI won’t bring you closer to your goals. You must start with the business problem and then determine if there is an AI-driven solution.

Still, you must be intellectually curious and willing to explore. While developing client-facing technologies demands a proven methodology and thorough testing for compliance and risk, you can learn much about AI by experimenting within your own four walls. So, try something new. Maybe you have 10 great AI-related ideas, you choose five to explore and one pays off. You can look at the result as four failures, or you can view it as a great success that helped you reach a goal while learning a lot along the way.

The key is to create a culture in which it is okay to fail. Part of that culture involves exposing your workforce to new technologies through training and participation in industry events. Move past the expectation that everyone must stay heads-down, focused on their narrow silo within your enterprise. Give people at every level of your organization the opportunity to pause, look up and look around at emerging technologies. They may just see opportunities you don’t see.

¹FIS, Global Innovation Research 2024 surveyed 2,000 firms across financial services and other industries in the U.S., U.K., Singapore, Hong Kong and Australia.