AI Disrupts Traditional Banking Revenue, Unlocking Fresh Growth

The panic around Meta’s Muse AI agent and its threat to bank profits obscures a more complex reality: artificial intelligence is simultaneously dismantling traditional revenue streams and creating entirely new ones for financial institutions willing to adapt.

When Meta released its Muse AI agent earlier this month, bank stocks took an immediate hit. The KBW Nasdaq Bank Index dropped 2.6% in a single day as investors imagined customers suddenly waking up to the realization that their checking accounts were earning pennies while money market funds offered substantially higher returns. Charles Schwab, JPMorgan Chase, and other financial giants watched their valuations shrink as the market priced in a future where AI agents optimize every dollar sitting in low-yield accounts.

The concern has merit. Banks have built profitable businesses on customer inertia, earning healthy margins from the spread between what they pay depositors and what they earn lending that money out. When customers leave cash sitting idle in checking accounts earning 0.01% while the bank lends it at 7%, that’s pure profit. Meta’s Muse, which can monitor account balances and automatically shift money to higher-yielding alternatives, threatens this comfortable arrangement.

But the fixation on this single threat misses the broader transformation underway. AI isn’t just helping customers squeeze banks for better rates-it’s fundamentally reshaping how banks operate, compete, and generate revenue. The institutions that survive this transition won’t be the ones that successfully defend their deposit margins. They’ll be the ones that recognize AI as an opportunity to rebuild their economic model from scratch.

Charles Schwab learned this lesson the hard way in 2023. As interest rates climbed, customers moved billions from their brokerage sweep accounts into money market funds. The company hadn’t anticipated the speed or scale of the shift. Suddenly forced to tap expensive short-term funding markets, Schwab watched its profits evaporate throughout the year.

This phenomenon, which analysts call “deposit sorting,” represents exactly what banks fear most about AI agents. Meta’s Muse connects to financial accounts through Plaid, monitors balances continuously, and can execute transfers without human intervention. The app offers personalized recommendations based on spending patterns and savings goals, making financial optimization effortless for users who previously lacked the time or knowledge to maximize their returns.

Bank of America analyst Ebrahim Poonawala captured the industry’s anxiety in a recent note to clients, writing that the rapid adoption of Meta’s app “establishes the opportunity” for margin compression. He’s watching deposit costs closely, knowing that if AI agents gain widespread adoption, banks will need to raise rates across the board to retain customer cash.

The timing couldn’t be worse for banks. Competition for deposits has already intensified as loan growth accelerated this year. Banks need funding to support new lending, but the personal savings rate is hovering near a four-year low. Meanwhile, digital banks-which operate without physical branches and can afford to pay higher rates-have been using social media to advertise their superior yields, attracting rate-sensitive customers who might have stuck with traditional banks in the past.

Citigroup recently rolled out a new savings rate initiative aimed at attracting more customer balances. PNC and Bank of America have launched similar programs. These moves signal that banks recognize they can’t simply ignore the pressure. The easy money from idle deposits is drying up, and AI agents will only accelerate that trend.

AI agents may erode banks’ profitable deposit margins by moving customer cash into higher-yielding alternatives, but that is only one side of the transformation. Banks can deploy AI to automate costly operations, improve fraud detection and underwriting, anticipate customer needs, expand wealth services, and generate new fee revenue. The institutions most likely to thrive will treat AI not merely as a cost-cutting tool or defensive response, but as a way to build more valuable and resilient customer relationships.

Here’s what the doomsday narrative overlooks: banks aren’t passive victims in this story. They’re also deploying AI aggressively, and the technology is opening revenue opportunities that dwarf the threatened deposit margins.

JPMorgan Chase CEO Jamie Dimon previewed a wealth management product called Smart Cash earlier this year. The tool will help customers optimize their cash allocation automatically-essentially offering the same service as Meta’s Muse, but within JPMorgan’s ecosystem. Bank of America already provides a corporate treasury platform that uses AI to help businesses forecast cash needs and manage liquidity. These aren’t defensive plays. They’re efforts to capture a larger share of customer assets by providing genuinely valuable services.

The banks understand something that the market panic missed: customers who engage more actively with their finances don’t just move money to higher-yielding accounts. They also discover investment opportunities, take out loans for major purchases, buy insurance products, and generally deepen their relationship with their financial provider. AI makes this engagement effortless, which means banks can scale these relationships in ways that weren’t possible before.

Research on AI adoption in banking shows the technology significantly improves profitability through two primary mechanisms. First, it reduces operational costs by automating routine tasks like fraud detection, compliance monitoring, and customer service inquiries. Banks are deploying AI for credit underwriting, pattern recognition in transaction monitoring, and regulatory reporting-all high-volume, pattern-driven tasks that consume enormous amounts of employee time.

Second, AI helps banks capture “money in motion” more effectively. By analyzing spending patterns and life events, AI systems can anticipate when customers might need a mortgage, car loan, or investment advice. Banks can reach out proactively with relevant offers, growing revenue without proportional increases in marketing spend.

McKinsey estimates AI could drive up to 20% net cost reductions across the banking industry as implementation matures. Another analysis suggests AI adoption could generate an additional $170 billion in profits for the global banking sector by 2028. These aren’t marginal improvements. They represent a fundamental shift in banking economics.

The market’s focus on deposit margin compression assumes banks compete primarily on price. But that’s never been entirely true, and it’s becoming less true as AI reshapes customer expectations.

Customers don’t choose banks solely based on who pays the highest interest rate. They choose based on convenience, trust, service quality, and whether the institution helps them achieve their financial goals. AI is transforming all of these factors simultaneously.

Consider fraud detection. AI models can analyze transaction patterns in real time, identifying suspicious activity before customers notice anything wrong. This isn’t just a cost-saving measure-it’s a trust-building service that makes customers more likely to consolidate their financial lives with a single provider.

Or take lending decisions. AI can expand credit access by identifying creditworthy borrowers who might be rejected by traditional underwriting models. This grows the addressable market while reducing underwriting costs and approval times. Customers get faster decisions, banks originate more loans, and default rates can actually decline because the AI models better predict repayment likelihood.

Wealth management provides another example. Traditionally, personalized financial advice was available only to high-net-worth clients who could justify the cost of a human advisor. AI makes personalized recommendations scalable to mass-market customers. Banks can offer goal-based planning, tax-loss harvesting, and portfolio rebalancing to millions of customers simultaneously, generating fee revenue from relationships that previously weren’t profitable to serve.

The institutions winning this competition won’t necessarily pay the highest deposit rates. They’ll be the ones that use AI to deliver such compelling value across the entire customer relationship that switching becomes unthinkable.

The central mistake is measuring AI’s impact on banking through deposit costs alone. Margin compression is real, but the more important question is whether a bank can convert better customer engagement, faster decisions, lower fraud losses, and scalable advice into durable revenue. AI will reward institutions that redesign their operating models around these capabilities, not those that simply automate existing tasks or defend yesterday’s spreads.

That doesn’t mean AI implementation is straightforward. Banks face real challenges that could derail the opportunity if not handled carefully.

Generative AI, in particular, presents risks that banks are still learning to manage. The technology can hallucinate false information, exhibit bias in decision-making, and make mistakes that are difficult to detect until they cause real harm. A recent analysis noted that generative AI accelerates research, drafting, and analysis but remains unreliable for final high-stakes decisions. Banks need robust governance frameworks and extensive testing before deploying these systems at scale.

Regulators are paying close attention. A recent speech by banking supervisors highlighted that AI adoption is compressing the timeframe for detecting and responding to cyber threats, challenging banks’ operational resilience. Supervisors are expanding their focus beyond internal AI governance to understanding how AI’s broader economic impacts might affect bank operations and business models.

There’s also a productivity paradox that hasn’t been fully resolved. Banks are deploying AI widely in support tasks like coding, customer service, payments reconciliation, and administrative work. But measurable impacts on cost reduction or revenue growth remain scarce. One analysis found that while AI reduces task completion times dramatically-cutting a one-hour job to twenty minutes, for example-banks haven’t yet redesigned their processes to capture that time savings as genuine productivity gains.

Time saved isn’t value created unless it’s redirected to higher-value activities. Most banks are still figuring out how to restructure workflows around AI capabilities. Until they do, the technology remains more promise than reality.

Bank of America’s Poonawala made an important point in his note to clients: “Until deposit costs rise faster than can be explained by rates or competition, the disruption thesis remains conceptual.”

He’s right that actual behavior change matters more than theoretical threats. But he’s measuring the wrong variable. Deposit costs might rise moderately as AI agents gain adoption, but that won’t be the defining metric of AI’s impact on banking.

The real transformation will show up in customer engagement levels, cross-sell ratios, operational efficiency metrics, and the ability of banks to compete profitably in markets they previously couldn’t serve. It’ll appear in fraud losses that decline while customer satisfaction rises. In loan portfolios that grow faster than funding costs. In fee revenue that offsets margin compression.

Huntington Bank provides a preview of this dynamic. The regional bank has seen deposit costs edge up to 1.88% due to competition and faster loan growth. But strong fee revenue growth-up about 30% year-over-year from payments, wealth, and capital markets segments-is more than offsetting the margin pressure. The bank uses AI and granular pricing strategies across 56 rate regions to optimize deposit costs while growing fee income.

That’s the actual story of AI in banking. Not a simple disruption narrative where AI agents steal deposits and banks suffer. A complex restructuring where traditional revenue sources face pressure while new opportunities emerge for institutions that adapt quickly.

The banks that thrive over the next decade won’t be the ones that successfully defend their deposit margins against AI-powered optimization. They’ll be the ones that deploy AI more effectively than their competitors, creating customer experiences and operational efficiencies that generate entirely new profit pools.

Meta’s Muse might help some customers earn better returns on their cash. But it’s also advertising to banks exactly what customers value and creating a roadmap for financial institutions willing to invest in similar capabilities. The disruption runs both ways, and the institutions that recognize that will come out ahead.