AI-Driven Digital Banking: An Advanced Overview for IT Leaders

How AI Is Reshaping Banking Operations

AI is altering the way banks operate, says Jerry Silva, program vice president for IDC Financial Insights. “Traditionally, it’s been used in areas like fraud detection and customer relationship management,” he says.

The introduction of generative AI (GenAI) in late 2022 marked a turning point, revolutionizing the experiences banks can provide customers and spawning new areas of innovation. Banks are now prioritizing education, adoption and risk management as they incorporate these tools into both customer-facing services and back-office processes.

“AI has overshadowed the cloud as the dominant technology focus,” Silva says.

Use Cases for AI in Banking

AI now touches day-to-day operations of banking from front office to back office, and GenAI is already in use for content and knowledge work. “There are institutions developing marketing materials and sales materials and doing knowledge management,” Silva says.

Simple use cases are having game-changing effects. For example, banks are automating internal support with chatbots for common requests. “Banks are now using GenAI to handle that password reset business instead of waiting for somebody from the help desk to call you,” he says.

DISCOVER: A new era of digital banking is powered by artificial intelligence.

Alyson Clarke, principal analyst for digital business strategy with Forrester, sees a mix of cost-cutting and growth opportunities for AI in banking. She notes that underwriting, fraud detection and risk management are areas where AI can streamline operations and improve accuracy.

“Underwriting and risk are big use cases in financial services,” Clarke says. “The real uplift comes from how AI helps firms drive deeper relationships with customers and become a more trusted adviser.”

How To Integrate AI in Legacy Banking Systems

Integrating AI into legacy banking systems remains a significant hurdle, particularly for institutions that have delayed infrastructure modernization. “Those banks that dragged their feet around legacy modernization are feeling the impact of not being able to use some of the AI capabilities that are coming out,” Silva says

The industry is making steady progress, and many financial institutions have upgraded their core platforms to support open application programming interfaces and more flexible connections. These upgrades make it easier to implement AI tools without the cumbersome integrations that characterized older infrastructure.

For smaller banks, engagements with managed service providers have become crucial, as AI introduces a level of technical complexity, risk and regulatory oversight that can be difficult for smaller institutions to manage internally. “More software providers are embedding AI into their systems,” Silva explains.

Adoption is happening fastest in front- and middle-office functions such as deposit and lending origination, customer relationship management, and underwriting. Core systems such as policy administration and trading platforms have been slower to evolve, but momentum is shifting.

Regulatory Compliance and AI

At the federal level, the emergence of AI has had limited impact on compliance efforts, as no specific AI regulations currently govern the U.S. banking sector. “I don’t anticipate for the next three years that there will be any new regulations specifically around AI in banking,” Silva says.

Agencies including the Federal Deposit Insurance Corp., Office of the Comptroller of the Currency and the Federal Financial Institutions Examination Council are exploring how to adapt existing rules, and states such as California have moved closer with privacy legislation, but nothing directly targets AI. However, that regulatory gap has not made banks complacent.

“Every institution I talk to is very cautious about how they’re implementing AI, particularly around use cases involving credit decision-making for example,” Silva says.

Institutions are proactively ensuring that their models remain explainable and transparent, treating these steps as best practices even without explicit rules.

EXPLORE: Artificial intelligence is empowering employees and IT leaders in the workplace.

Benefits and Challenges of AI in Digital Banking

While AI offers major benefits, Clarke cautions that many banks risk repeating mistakes made during digital transformation by focusing too narrowly on efficiency. “Most banks still struggle with siloed data and outdated systems, which limits their ability to extract insights from AI tools,” she says.

Without strong data strategies, firms risk using AI to speed up existing processes without changing business models or growing revenue.

The key, Clarke says, is shifting the conversation away from “How do we use AI?” to “What are we trying to achieve, and how can AI help us get there?”

The Future of AI in Banking in 2026 and Beyond

Looking ahead, Silva sees banking AI investment shifting from cost savings to innovation. “Almost 75% of the spend is around automating business processes or automating task processes,” he says.

By 2028, he expects more than half of AI budgets to fund new products and services across lending, wealth management and payments. “You invest in automation, you bring that cost basis down,” Silva says. “You can use those funds to boost innovation.”

Clarke says the next few years will likely see banks move beyond hype and start using AI more effectively in core areas such as customer retention and deposit management. She identifies predictive AI as an underused capability that could help institutions orchestrate customer journeys and strengthen loyalty if applied strategically.

While many banks are rushing toward the latest advances, such as agentic AI, she says even established tools have untapped potential to improve marketing, personalize interactions and reinforce trust. “Over the next few years, we are definitely going to see more firms leveraging AI for fraud, risk and underwriting,” Clarke says.

Sliva points to the rise of personal AI agents , digital clones that can manage emails, calendars, shopping and even banking on a customer’s behalf.

“The technology is already possible, but unresolved questions of liability and trust will determine how quickly it comes to market,” he says.