By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
While the 'soft switching' phenomenon is concerning, it's not yet catastrophic. Banks are caught between cost pressures and service expectations, and the real issue is how they execute AI deployment and manage customer expectations. Regional banks face risks, but megabanks have switching costs that limit real defection.
Risk: Permanent reduction in deposit beta efficiency and higher cost of funding during liquidity tightens (Gemini), AI eroding mortgage refi cross-sell (Grok), and deposit beta efficiency breaking in 2026 when promotional cohorts mature (Claude)
Opportunity: Pairing AI-assisted efficiency with robust human escalation and product design (ChatGPT)
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Pope Leo got hung up on by his own bank — why so many Americans are now quietly moving their money elsewhere
Kit Pulliam
5 min read
"Could you imagine being known as the woman who hung up on the Pope (1)?"
Father Tom McCarthy, a personal friend of Pope Leo XIV, said this while sharing a story about the Pope's struggles dealing with his bank's customer service.
The Pope was trying to get the phone number changed for his bank account, which was held by a bank in his hometown of Chicago. Despite answering the security questions correctly, the customer service representative said he would have to come in to change the number, which he couldn't do.
Dave Ramsey warns nearly 50% of Americans are making 1 big Social Security mistake — here’s how to fix it ASAP
"Would it matter to you if I told you I'm Pope Leo?" he said. The representative hung up on him.
Luckily for the Pope, he knew a guy who knew a guy. He was able to escalate the issue to the bank president and get his number changed; understandably, they didn't want to lose the Pope to another bank.
Pope Leo isn't alone in his customer service struggle — although most people don't have his connections to lean on. The J.D. Power Retail Banking Satisfaction Study shows that customers are beginning to show dissatisfaction with their banks' customer service options (2).
Here's what people are dissatisfied about — and what banks could do to fix it.
People are starting to "soft switch" their bank accounts more often
The 2026 J.D. Power study shows that customer satisfaction rates sharply declined in the second half of 2025. It says this decline shows "growing strain in the customer experience with phone, branch, online and automated customer support channels."
Customers are showing their dissatisfaction by slowly transferring money away from their primary account to another bank. 20% of retail customers are "soft switching" their accounts like this, up from 17% in the previous year.
"Cracks are emerging at key points in the customer journey, which are opening the door to quietly establish new accounts with other institutions and gradually shift funds away from their primary bank," says Jennifer White, senior director of financial services intelligence at J.D. Power.
This could be because banks are increasingly using AI as part of their customer service approach, something that consumers are only fine with in specific circumstances.
According to a T.D. Bank survey, two-thirds of people were open to AI being used behind the scenes for things like fraud detection (3). But over 80% of people would prefer to be able to reach a human when they call the bank for customer support.
A 2023 report from the Consumer Financial Protection Bureau found that, even back then, chatbots were beginning to replace traditional call centers as banks' first-line customer support solution (4). This frustrated people, who were left to deal with nonhuman chatbots with no way to escalate.
"What is worse is there is not way [sic] to contact a person who can actually resolve the situation," said one CFPB complaint.
Even if there is a way to contact a human customer service representative, it can be complicated to figure out how. Some banks require you to enter your account information before moving forward; others have complicated and hard-to-parse phone trees that could lead to you being shuffled between customer service representatives when trying to find the right person to speak to (5).
Many banks made customers sit through long wait times before they spoke to someone. If you didn't reach the right person, you'd have to get back in line and wait again.
Banks frequently use chatbots or other AI services because they're cost effective, quick and good at handling simple repetitive tasks (4). If banks want to improve their customer service scores, they don't have to get rid of AI as an option entirely: Almost half of people are open to using AI banking assistants for simple things like paying bills and setting alerts (3).
But banks might want to also offer well-staffed customer service lines for the customers who aren't comfortable using those tools, as well as for more complicated issues or financial questions.
White says one of the best ways banks can increase their customer service scores is to make it easy to solve any problems that come up. She says that banks that score high on customer service overall "are scoring strongly on resolving problem friction well (6)."
"Because sometimes when you have a problem, you know, a good resolution can actually result in higher customer experience satisfaction scores than never having the problem to begin with," she says.
Banks that successfully reduce problem friction will have happier customers; they'll also lose fewer customers to soft switching.
Four leading AI models discuss this article
"The rise of 'soft switching' indicates that regional banks are losing their moat as primary transaction hubs, which will eventually erode their access to low-cost core deposits."
The 'soft switching' phenomenon highlighted by J.D. Power is a canary in the coal mine for regional banks, specifically those lacking the scale to invest in high-end human-centric CX (customer experience). While the article frames this as a service failure, it is actually a margin-preservation strategy. Banks are trading customer satisfaction for reduced headcount costs to offset compressed Net Interest Margins. If this trend persists, we will see a widening valuation gap between 'digital-first' giants like JPM or BAC, which can afford both AI and human support, and smaller regional players (like KEY or ZION) who risk becoming mere liquidity conduits for customers who keep their primary relationships elsewhere.
The cost of human support is a massive anchor on ROE; banks that successfully automate away the 'human' layer may actually see superior long-term profitability despite higher churn rates among low-value retail depositors.
"Soft switching's 3pp rise to 20% highlights deposit stickiness more than fragility, enabling AI cost savings to sustain bank profitability."
The Pope anecdote grabs headlines, but J.D. Power's data reveals modest strain: soft switching rose just 3pp to 20%, far from a deposit exodus amid FDIC-insured stability and banks' 4-6% YoY retail deposit growth through 2025 (FDIC Q3 data). AI-driven cost cuts (20-30% opex savings per industry benchmarks) bolster NIMs and ROEs for majors like JPM (12% ROE), BAC, outweighing service gripes. Regionals (e.g., FITB, HBAN) face higher friction risk due to thinner margins. Fintechs like SOFI, NU nibble share but struggle with scale. Smart banks resolving 'problem friction' turn this into loyalty gains.
If escalating AI frustrations lead to viral backlash or regulatory mandates for human staffing (as CFPB complaints hint), deposit outflows could accelerate beyond 20%, hammering regional banks' funding costs amid still-high rates.
"J.D. Power's 3-point uptick in soft-switching is real friction, but the article provides no data on actual deposit outflows or which banks are winning — making it impossible to assess whether this is a structural threat or a service-quality fix opportunity."
The article conflates anecdote with trend. Yes, J.D. Power shows satisfaction declining and 20% soft-switching (up from 17%), but that's a 3-percentage-point move — meaningful but not catastrophic. The real issue: banks are caught between cost pressure and service expectations. AI deployment is rational economically; the complaint is about execution (no escalation paths). The Pope story is clickbait masking a legitimate operational problem. What's missing: whether soft-switchers actually *leave* or just hedge, and whether this translates to deposit flight or just account fragmentation. Regional banks and community lenders may gain share, but megabanks have switching costs (direct deposit, mortgage relationships) that limit real defection.
Soft-switching may signal healthy competition and account diversification rather than bank weakness — customers keeping money at multiple institutions could reflect rising rates on savings accounts elsewhere, not dissatisfaction. And 20% of retail customers soft-switching doesn't mean 20% are leaving; many may maintain both accounts indefinitely.
"The real alpha is in customer-experience excellence: banks that blend AI-driven efficiency with reliable human support will improve retention and cross-sell, not just reduce costs."
Article frames AI-driven customer service as a growing customer churn risk for U.S. banks, but the signal is more nuanced. The 20% 'soft switching' stat suggests movement, yet not a free-fall of deposits; real outflows and loan growth dynamics matter more. The piece leans on anecdotes (Pope, calls) and sentiment surveys; it omits margin, funding costs, and the durability of deposit bases in a higher-rate regime. The opportunity for banks is to pair AI-assisted efficiency with robust human escalation and product design; the risk is overreliance on chatbots without resilience or proper pricing power. The bigger alpha: CX excellence can improve retention and cross-sell, not just trim costs.
But the headline risk may be overstated: 'soft switching' could reflect promotional churn rather than durable deposit loss. AI can cut costs only if outages and fraud risks are managed.
"Soft-switching represents a structural degradation of deposit stickiness that will force regional banks to pay higher funding costs."
Claude, you’re missing the second-order effect of 'account fragmentation.' If 20% of customers are hedging by keeping balances elsewhere, the 'sticky' core deposit base—the lifeblood of regional bank valuation—is structurally eroding. This isn't just competition; it’s a permanent reduction in deposit beta efficiency. When liquidity tightens, these 'hedged' deposits will be the first to flee to higher-yielding digital-first treasuries, leaving regionals like ZION or KEY with a higher cost of funding and compressed NIMs.
"Soft switching fragments promo-chasing savings, not core sticky deposits, preserving regional funding bases."
Gemini, your fragmentation thesis ignores deposit composition: J.D. Power implies soft switching hits low-balance retail savers chasing 4.5-5% promo rates elsewhere, not core operating deposits (60%+ of regional funding per FDIC). This boosts average deposit yields without raising funding costs structurally. Unflagged risk: if AI erodes mortgage refi cross-sell (5-7% of non-interest income), that's the real ROE killer for KEY/ZION.
"Promotional deposit churn is a lagged problem, not a current one—regionals' funding cost crisis arrives when rate-chase deposits mature, not when customers open them."
Grok's deposit composition rebuttal is stronger than Gemini's fragmentation thesis, but both miss the timing risk. Yes, soft-switchers are rate-chasers hitting promo accounts—but those promos expire in 12-18 months. When they do, regionals face a choice: match rates (margin compression) or lose deposits (funding cost spikes). The real test isn't 2025; it's Q3-Q4 2026 when promotional cohorts mature. That's when deposit beta efficiency actually breaks.
"Core deposits aren’t a shield; AI-driven churn can push regionals into higher-cost, non-core funding during stress, compressing NIM beyond the 60% core-deposit baseline."
Grok, your 60% core-deposit baseline may understate fragility. Even with sizable core deposits, 20% soft-switching likely concentrates churn into the higher-cost, thinner-margin segments first, and the frag goes beyond mix—brand trust and AI-driven customer journeys can deteriorate quickly under stress. In a downturn, non-core funding (brokered deposits) can spike, pushing funding costs higher for regionals and compressing NIM more than your 60% figure implies.
While the 'soft switching' phenomenon is concerning, it's not yet catastrophic. Banks are caught between cost pressures and service expectations, and the real issue is how they execute AI deployment and manage customer expectations. Regional banks face risks, but megabanks have switching costs that limit real defection.
Pairing AI-assisted efficiency with robust human escalation and product design (ChatGPT)
Permanent reduction in deposit beta efficiency and higher cost of funding during liquidity tightens (Gemini), AI eroding mortgage refi cross-sell (Grok), and deposit beta efficiency breaking in 2026 when promotional cohorts mature (Claude)