Clootrack is now officially available as a ChatGPT Plugin
Install now
Five shifts are changing customer experience across U.S. banking, insurance, and wealth management in 2026: primary relationships fragmenting without a formal switch, aggregate satisfaction scores masking sharp divergence underneath, impersonation fraud targeting firms' own alert channels, consumers splitting tasks between machines and people by consequence, and product discovery moving into AI-generated answers.
Customer experience in financial services is the sum of every interaction a customer has with a bank, credit union, insurer, or wealth management firm, across branches, mobile apps, contact centers, advisers, and now third-party AI assistants that answer questions on the customer's behalf. Measuring it well increasingly depends on Voice of Customer analytics rather than survey scores alone.
Most of what passed for a financial services CX trend two years ago has either resolved or stopped being a trend. Digital adoption is settled. Self-service is table stakes.
What is unsettled is harder to see. Every shift below shares one property: the customer's behavior, expectations, or exposure is changing faster than the metrics institutions use to manage the relationship. This analysis covers U.S. retail banking, property and casualty insurance, and wealth management, drawing on 2025 and 2026 research from J.D. Power, the American Customer Satisfaction Index, the Federal Trade Commission, the American Bankers Association, and Santander US.
Across all three sectors, customers are showing less willingness to treat one institution as the default for every financial need.
In banking, the J.D. Power 2026 U.S. Retail Banking Satisfaction Study, based on 107,059 customers and released in March 2026, found the average checking customer now holds three deposit accounts at different institutions. J.D. Power calls this soft switching and frames the risk as gradual erosion of share of wallet rather than immediate attrition. It concentrates in the customers institutions most want to keep: those under 40, the affluent, and the financially healthy were all likelier than average to move money away.
Insurance shows the same instinct through shopping. J.D. Power's 2026 U.S. Insurance Shopping Study found shoppers now get an average of 3.5 quotes, the highest in the study's 20-year history, even as the share shopping for auto insurance eased from 57% to 53%. Fewer people shop, and those who do compare more carriers than ever.
Wealth management is fragmenting by brand. The J.D. Power 2026 U.S. Investor Satisfaction Study found that two of the top three ranked brands for do-it-yourself investor satisfaction are fintechs. Among investors under 40, brands including SoFi and Ally are seen as more innovative than established firms and equally trustworthy, with the share rating fintechs trustworthy up seven percentage points in a year.
Traditional attrition metrics can miss this stage, because an account, policy, or investment relationship can stay formally active while share of wallet shifts elsewhere. Detecting it depends on competitive benchmarking against the institutions competing for the same customers, not against a national average.
Industry averages are concealing more than they reveal. The ACSI Finance Study 2026, based on 14,210 surveys collected from U.S. customers through 2025, put banks overall at 80 on a 100-point scale, unchanged for a third year. Underneath that flat number, regional and community banks held at 83 while super regional banks fell 3% to 77, with national banks steady at 79. Credit unions slipped a point to 78, continuing a seven-year reversal against banks.
Retail banking shows the same effect across time. Overall satisfaction rose two points to 657 on a 1,000-point scale, while satisfaction fell sharply in the second half of the year across phone, branch, online, and automated channels. An annual score can move in the opposite direction from the experience underneath it.
Insurance is flat at a low level rather than a comfortable one. The J.D. Power 2026 U.S. Auto Insurance Study found satisfaction unchanged at 644, matching the prior year's record low, even as price satisfaction improved three points. Rates eased, but satisfaction did not. J.D. Power identifies the inability to deliver seamless interactions across channels as the single biggest drag.
Fraud has become a customer experience problem because criminals now impersonate financial institutions' own communications. The Federal Trade Commission reported in June 2026 that consumers lost about $16 billion to fraud in 2025, the highest total on record and an increase of about 25% over 2024. Imposter scams accounted for $3.5 billion and were the most reported fraud category that year.
The mechanism should concern every CX leader. The FTC notes that some of the costliest impersonation scams begin with a fake security alert, often appearing to come from a bank, after which the customer is persuaded to move money to protect it. Within roughly $1 billion in reported losses to business impersonators, the highest losses were attributed to criminals posing as banks.
The implication for CX is significant. Proactive alerts and urgent outbound contact were designed to build trust, but those same patterns can now be imitated by criminals. Every legitimate alert therefore has to compete with fraudulent communications for the customer's belief. A Credit One Bank survey reported in September 2026 found more than two in five U.S. consumers had encountered an AI-powered scam, personally or through someone they know.
Trust is the dimension J.D. Power ranks first in importance in retail banking, and one of seven evaluated in wealth management. Alert design, channel authentication, and post-incident recovery are now shared territory between CX, fraud, and risk.
The research suggests customers are increasingly separating financial tasks by consequence, not simply by convenience. A Santander US survey of middle-income consumers published in August 2026 found 90% would rather speak with a knowledgeable person than use AI for major financial decisions, and 89% want to keep access to a branch staffed with people who can help. The tasks they least want to handle online were resolving complex issues, speaking with a banker, and discussing financial decisions.
Routine banking has gone the other way. A national survey conducted by Morning Consult for the American Bankers Association in October 2025, among 4,403 U.S. adults, found 54% mainly use a mobile app, 22% online banking, 9% a branch, 6% ATMs, and 4% the phone. Mobile has led for a sixth consecutive year.
Wealth management shows automation and human advice can be complements. Among DIY investors using a robo advice platform, 17% say they are definitely likely to work with an adviser within the next year, against 4% of those who do not, rising to 28% among affluent users. J.D. Power characterizes robo advice as a gateway to human advice rather than a replacement.
Insurance shows the cost of getting the handoff wrong. Only 58% of auto insurance customers say they fully understand their policy, and 46% used multiple channels to reach their insurer last year. Both gaps raise the stakes when customers have to move between channels to resolve a single question. Contact center conversational analytics is where those breakdowns can become visible.
Many U.S. consumers now research financial decisions through an AI assistant before reaching an institution's own channels, and in one sector the effect on switching is measurable. The Santander research found more than four in 10 consumers have used AI for a financial task, with leading uses including finding ways to save money and understanding rates, fees, and terms.
Insurance provides the harder number. J.D. Power found 32% of auto insurance shoppers used AI tools during their search, and those shoppers were more than 1.3 times as likely to switch insurers. A similar share, 33%, found the AI content unhelpful.
The risk is that when an assistant explains a product poorly, the institution may absorb the consequence without ever seeing the interaction.
Traditional dashboards do not capture this. An institution can hold its rankings, app ratings, and satisfaction scores while losing the moment where the customer decides. The starting point is to establish whether it appears, and appears accurately, in AI-generated answers for its core rate, fee, and coverage queries. Incorrect product terms in a generated answer can become a marketing problem and a risk that compliance teams need to assess.
Detecting these shifts requires evidence that can surface between measurement cycles. Periodic survey scores remain useful for benchmarking, but they can miss changes that emerge between surveys. The earlier signals can appear before the formal metric moves: balances can leave before accounts close, policy comparisons can happen before renewals lapse, and trust in alerts can erode before engagement metrics move.
Some of the earliest signals can appear in unstructured customer language: contact center transcripts, app store reviews, complaint narratives, claims correspondence, chat logs, and social conversation.
Clootrack's client work in banking shows this in practice. One retail bank was already collecting more than 60,000 customer responses a month across NPS surveys, email, chatbot, app store reviews, and its website, but the score carried no explanation. The instructive finding was how fast the priority order moved underneath it. Mobile banking ranked first one month and fifth the next once fixes landed, complaint resolution took the top position after that, and interest rate rose to first once the bank adjusted its rates. The full account is in the case study on how a banking major reduced user churn.
One finding also means different things to different leaders, which is why a single recommendation fails. Rising dispute friction is a workflow defect to a product leader, a capacity question to an operations leader, a fraud-pattern signal to a risk leader, and a message-authentication problem to a marketing leader.
The largest identifiable gap in the current data is a conversation that is not happening. Just 51% of advised investors under 40, and 39% of clients aged 40 and older, say their adviser has discussed the elements needed for a future wealth transfer. Only 18% say their adviser has met with or suggested meeting additional family members. As wealth passes to younger investors who increasingly view fintechs as equally trustworthy, relationships that never extended beyond the account holder may be more vulnerable during the transfer.
Underneath that is a pace mismatch: priorities can reshuffle within a quarter while customer intelligence still arrives on a reporting cycle built for a slower market. Clootrack's MCP integration provides governed access to churn, loyalty, and Voice of Customer data from inside the AI tools leaders already use, so a question about why complaints rose in a product line can be answered during the meeting.
Financial services CX has entered a period where reported numbers and underlying reality can diverge for months at a time. Bank satisfaction holds while deposits leave. Insurance satisfaction holds flat while shoppers collect more quotes than ever. Younger DIY investors increasingly view fintech brands as equally trustworthy. The institutions that navigate 2026 well will stop treating aggregate scores as the whole picture and start reading what customers actually say, in the channels where they say it first.
See how financial services teams surface early customer signals from contact centers, reviews, and complaints before they reach a quarterly score: Clootrack platform and method.
The five major U.S. trends examined here are primary relationships fragmenting without a formal switch, aggregate satisfaction scores masking divergence underneath, impersonation fraud aimed at institutions' own alert channels, consumers splitting tasks between machines and people by consequence, and product discovery moving into AI-generated answers.
They are more comfortable with AI for routine tasks than consequential ones. Santander US research published in August 2026 found 90% of middle-income consumers would rather speak with a knowledgeable person about major financial decisions, even though more than four in 10 have already used AI for some financial task.
AI use is associated with a higher likelihood of switching. J.D. Power found in 2026 that 32% of auto insurance shoppers used AI tools during their search, and those shoppers were more than 1.3 times as likely to switch insurers. A third also found the AI content unhelpful.
Fraud can undermine confidence in legitimate alerts, because scammers increasingly imitate the same communication patterns institutions use. When criminals send fake security alerts that lead customers to move money, proactive outreach becomes harder for customers to verify. Fraud is now a design constraint on customer communication rather than a separate security concern.
Yes, for a narrower and more valuable set of tasks. Mobile is now the most preferred banking channel, but consumers still want branch access for complex issues and financial guidance. Branches also work as a credibility signal: 83% of Santander survey respondents said they had more confidence in a digital banking provider that maintained physical branches.
Soft switching is the pattern in which a customer opens accounts at other institutions and gradually shifts funds away from their primary bank without closing the original relationship. It is difficult to detect through closure-based attrition metrics because the account stays open while its share of the customer's money falls.
By analyzing unstructured customer language at scale. Contact center transcripts, app store reviews, complaint narratives, and claims correspondence can provide more granular signals between survey cycles, including the language customers use to describe emerging problems.
Share :
Analyze customer reviews and automate market research with the fastest AI-powered customer intelligence tool.
.webp)