Banks and Retailers Both Discovered Brand Loyalty in 2026. Their AI Strategy Is Engineered to Undermine It.
On the same September Thursday, McKinsey published a strategy piece arguing that AI personalization helps banks build customer loyalty, and Roger Dooley's Neuromarketing reminded its readers that paying for something literally activates pain centers in the brain. Neither article mentioned the other. They should be read together, because they describe the same mechanism from opposite ends, and the conflict between them reveals something most brands running loyalty programs have not yet worked out.
The McKinsey piece is confident. AI-powered personalization can help banks compete against digital-first rivals by delivering hyperpersonalized engagement at scale. The playbook is familiar: use AI to reduce friction, anticipate customer needs, smooth the transaction experience, and convert satisfied customers into loyal ones. It is framed as a competitive necessity because digital-first fintech rivals have already deployed this approach and traditional banks need to catch up.
The neuromarketing piece makes a different observation. Spending money activates pain centers in the brain. Handing over cash produces a measurable neural response that card payments reduce and contactless payments reduce further. The entire arc of payment technology innovation has been to decrease this response. Tap-to-pay exists specifically because friction-free payment produces higher spending. Less pain means more transactions. More transactions means more revenue. That is the product logic.
The problem with that logic, when you apply it to loyalty, is that pain-of-paying is not just a barrier to be engineered away. It is the signal that creates psychological commitment.
Why Brand Loyalty Suddenly Matters to Every Major C-Suite
The timing of all this is not coincidental. Brand loyalty is having its moment. YUM! Brands CEO Chris Turner used Q2 2026 earnings to argue that loyalty creates better experiences, stronger engagement, and more powerful demand-driving tools for franchise partners. United Airlines, Kohl's, Ulta Beauty, and Estée Lauder made similar arguments in their own earnings calls. After years of pundits dismissing loyalty programs as expensive noise, company after company is repositioning loyalty as a core driver of business performance.
The shift is real and the underlying economics support it. A loyal customer costs less to retain than a new customer costs to acquire. In categories where switching is infrequent (banking, airlines, grocery, beauty retail), a loyal customer's lifetime value compounds in ways that customer acquisition math dramatically undercounts. Loyalty programs also generate first-party data at a time when third-party cookies are degrading and digital advertising is facing structural pressure.
What none of these earnings calls address is where loyalty actually comes from. The standard assumption is that loyalty follows from satisfaction. Make the product better, make the experience smoother, make the interface more intuitive, and satisfaction increases. Satisfied customers stay. That is the model. And it is partly wrong in a way that matters a great deal when you are deploying AI to execute it.
The Commitment Signal That Friction Creates
Behavioral economists have known for decades that effort and commitment are intertwined. The IKEA effect (named for the furniture retailer that sells you products requiring hours of assembly) demonstrates that people value things they have worked to obtain more than equivalent things handed to them. This is not irrational. Effort is information. If you chose to do something difficult, your choice reveals genuine preference in a way that effortless choices do not. The brain uses past behavior as evidence about current preferences, and difficult past behavior is stronger evidence than easy past behavior.
Payment friction works the same way. When paying for something hurts (even mildly, even neurologically), the payment becomes information about commitment. You spent real money. You felt it. That signal updates your internal model of how much you value the thing you bought. Cash payments create stronger psychological ownership than card payments. This is not speculation. It has been replicated across purchase categories and cultures.
Contactless payments, AI-generated payment recommendations, and automated bill management tools are all designed to reduce the salience of money leaving your account. This is exactly what makes them good for driving transaction volume. And it is exactly what makes them a poor tool for building loyalty, because they remove the commitment signal that payment pain encodes.
The McKinsey AI personalization framework for banking works by making banking feel effortless: personalized offers appearing at the right moment, automated savings optimizations, frictionless cross-product bundling. Every one of these features is genuine improvement from a user experience perspective. None of them generates the kind of commitment signal that a customer who consciously set up their mortgage, their retirement account, and their checking with the same institution over ten years of deliberate decisions has encoded in memory. One customer has a relationship. The other has preferences that are one competitor offer away from reallocation.
What the Google Antitrust Outcome Tells You About Behavioral vs. Structural Loyalty
The Google antitrust ruling that landed this week clarified something important. Google dodged a forced breakup, and industry experts are split on whether the behavioral remedies the court imposed will actually restore competition. The majority view among technologists and economists seems to be that the alternative remedies are misdirected because Google's market power is behavioral, not structural.
Breaking up Google's ad exchange or forcing divestiture of its publisher server would address the structural conditions that allow Google to favor its own properties. But Google's search dominance is not primarily structural. It is behavioral. Users choose Google. Default settings help, but people who change their default to Bing or DuckDuckGo tend to change it back. Google's position persists because users have built habits and heuristics around it that function independently of its ownership structure.
Brand loyalty is the same distinction. Loyalty programs and AI personalization tools address the structural conditions for retention: points, offers, personalized engagement, frictionless experience. These are real and they help with churn. But the deepest form of loyalty, the kind that persists through competitor offers and product missteps, is behavioral. It was formed through repeated, deliberate choices that encoded commitment. You cannot engineer your way to that outcome by making every interaction effortless. The effort is part of how it was built.
This is the pattern previous STI analysis of agentic commerce covered: agentic AI that removes human decision-making from routine purchases also removes the human decision signal that makes those purchases loyalty-building. An AI agent that auto-renews your subscription, optimizes your subscription portfolio, and switches you to a cheaper competitor when pricing changes is not a loyalty machine. It is a churn-optimization machine running on your behalf rather than the company's.
The Roth Conversion Window and What It Reveals About Financial Commitment
The Kiplinger piece on soft retirement and Roth conversions is ostensibly about tax strategy. But it describes something more interesting: a psychological commitment event. Phasing out of full-time work creates a transition window where deliberate, complex financial decisions compound over years. People who execute Roth conversions during this window are not making passive choices. They are making effortful, high-stakes, friction-full decisions that require understanding tax brackets, conversion windows, penalty rules, and long-term projections.
The bank or financial advisor that guides someone through this process does not just capture the assets. It captures the memory of the effort. That relationship has a fundamentally different commitment structure than a relationship where an AI automatically optimized the tax situation in the background with zero user engagement. Both outcomes might be financially equivalent. The first one builds loyalty. The second one builds convenience dependency, which looks identical until a competitor offers a marginally better interface.
This is the original contribution that connects the pieces the individual articles each miss: AI personalization tools in banking are reducing payment salience and decision effort precisely in the interactions that historically created deep loyalty. Optimized for conversion rate and user experience metrics, they are competent at what they measure and blind to what they erode.
What Brands Running Loyalty Programs Should Actually Do
None of this argues that banks should make their apps deliberately worse or refuse to adopt AI personalization. The competitive reality McKinsey describes is real. Banks that ignore AI personalization will lose share to banks that deploy it. The question is whether "deploy AI personalization to build loyalty" is the right strategic framing, or whether it is a measurement category error.
The more useful frame is what STI has documented around brand trust as an operational property: trust and loyalty are outcomes of specific interactions, not of satisfaction scores in the aggregate. The interactions that create durable loyalty are ones where customers make visible, deliberate choices with real consequences. A customer who chooses to consolidate their financial life at one institution after evaluating alternatives has created a loyalty signal. A customer who stayed because the app kept getting better notifications has not.
The AI personalization playbook that McKinsey describes is genuinely good at the latter and has nothing to say about the former. What it cannot do is create the conditions for the deliberate commitment decisions that build loyalty durable enough to survive a competitor's offer. For that, you need friction. Not the bad kind, the kind that comes from poor interfaces and confusing fees, but the intentional kind: the moment where a customer chooses you, knows they are choosing you, and encodes that choice as identity.
YUM! Brands and the other companies suddenly talking about loyalty economics are right that loyalty changes the math. They have not yet answered what changes the loyalty. Behavioral science has the answer. It is the one thing their AI personalization strategy is designed to eliminate.
The companies that figure this out first will not be the ones with the smoothest customer journeys. They will be the ones that have learned to distinguish between friction that signals poor design and friction that signals genuine commitment, and have the discipline to preserve the latter while eliminating the former. That is harder to build than a personalization model. It is also harder to copy.
If you are thinking through what this means for your own brand strategy, STI's research team publishes ongoing analysis at smarttechinvest.com/research.