Avery Miller, Visa's VP of Loyalty Solutions, recently said what the loyalty industry has been avoiding: the credit card can't carry loyalty alone anymore.

In a detailed interview with PYMNTS, Miller laid out the pressure points. Interchange margins are compressing. Card-funded rewards can't scale. Every bank's loyalty program looks identical to the customer. Consumers are enrolled in eighteen or more programs but actively engage with roughly half. The economics underneath card-funded loyalty are tightening, and the industry built on top of that funding model is starting to feel it.

Miller's proposal: expand what gets rewarded beyond just card spend — bill pay, deposits, wallet provisioning, digital sign-on — and fund those rewards through merchant-funded offers rather than pure interchange margin. Add contextual AI delivery so the right offer reaches the right person at the right moment. Make each program smarter, more personalized, more relevant.

It's a meaningful upgrade. And it misses the actual problem.

What Visa Is Actually Saying

It's worth being precise about Miller's argument, because the LinkedIn commentary around this article largely misread it.

Visa isn't advocating away from payments as the capture point. They're advocating for expanding which behaviors count as loyalty-worthy while keeping the payment infrastructure as the primary delivery channel. The transaction isn't going away — Miller explicitly noted that in a world of AI agents, the fight will happen for every single transaction, with issuers competing repeatedly for share of mind rather than just share of wallet.

That's actually more transactional, not less. They're making the transaction smarter with contextual delivery. Not questioning whether the transaction is the right unit of loyalty in the first place.

The merchant-funded offer model is also revealing. When interchange margins compress, someone else has to pay for the reward. Merchants will do it — in exchange for access to new customers, larger baskets, and better conversion data. That's a real business model innovation. But it changes who funds the closed loop. It doesn't open it.

The 18-Program Confession

The most important data point in the interview wasn't a strategy recommendation. It was an admission.

Visa's own VP confirmed the stat that Resonance's research has been tracking for months: consumers are drowning in loyalty programs that don't talk to each other. Half those memberships are dead weight. The differentiation between programs has collapsed. Consumers aren't struggling because program #7 isn't personalized enough. They're struggling because programs #1 through #18 exist in total isolation from each other.

Visa's response to this: make each of those 18 programs individually smarter. Better data. Better timing. Better merchant-funded offers. If program #7 gets 15% more personalized, the consumer will engage more with program #7.

Maybe. But that doesn't solve the architectural problem. It optimizes within a broken structure.

The Closed-Loop Upgrade vs. The Open-Loop Question

This is the distinction that matters. Visa is proposing a closed-loop upgrade — make each isolated program better. Resonance is proposing an open-loop architecture — make the value portable across programs.

These aren't competing ideas in the same category. They're answers to different questions. Visa is asking: "How do we make card-linked loyalty programs better?" That's a legitimate question with real business impact. Merchant-funded offers, expanded behavioral triggers, AI-powered contextual delivery — all of that makes the existing infrastructure more effective.

The question it doesn't ask: "What would it take for the value a customer earns in one place to mean something everywhere else?"

That's not a personalization problem. It's a network problem. And you can't solve a network problem by making silos more sophisticated.

The Demand Signal Visa Confirmed

Nine in ten loyalty members want to redeem rewards across multiple brands. That stat comes from ValueDynamix's 2024 research, and Visa's own admission about program fatigue confirms the underlying pressure.

Consumers aren't asking each of their 18 programs to get smarter. They're asking why these programs don't interoperate. Why the effort they put into Brand A is invisible to Brand B. Why every new relationship starts from zero.

This isn't a fringe demand. Acquiring a new customer costs five to twenty-five times more than retaining an existing one. A 5% increase in retention correlates with 25-95% profit growth. Communities increase customer retention by 40%. The economic case for keeping customers engaged across a network is overwhelming.

Yet 44% of companies still prioritize acquisition over retention. The loyalty industry responds to this by adding another closed-loop program to the pile. Consumer gets membership #19. Same architecture. Different logo.

Where This Gets Interesting

Miller mentioned AI agents repeatedly — and this is where the real implications live. In an agentic commerce world, AI agents handle purchasing decisions on behalf of consumers. They compare options, evaluate rewards, and route transactions. Miller sees this as issuers competing for every single transaction through the agent layer.

But here's what that means for loyalty architecture: most loyalty programs are UI-dependent. They require a human to open an app, navigate a portal, scan a card. An AI agent can't do that. The program becomes invisible to the agent, and invisibility means irrelevance.

The programs that survive agentic commerce are the ones with machine-readable infrastructure — APIs that agents can query, balances they can check, redemptions they can execute programmatically. That's not a feature you bolt onto a card-linked program. It's a fundamentally different architecture.

What Resonance Takes From This

Visa validating the cracks is useful. Margin compression, program sameness, engagement drop-off, agent disruption — these are the exact pressure points that create demand for open-loop infrastructure.

But validation of the problem is different from alignment on the solution. Visa's answer keeps value inside closed loops with better tooling. Resonance's thesis is that the value should flow between participants through an open network — brands fund rewards on their terms, customers earn with one and redeem across all, effort compounds instead of expiring.

The economics support this. Open networks with genuine value persistence see redemption rates of 60-70%, compared to 15-25% in closed programs. Post-redemption, customers who actually use their rewards increase spending by 83%. Non-redeemers churn at 2.3x the rate of redeemers.

The industry profits from silence — from breakage, from unredeemed points, from the $360 billion in loyalty liabilities sitting on corporate balance sheets worldwide. Resonance profits when people actually use what they've earned.

Visa sees the cracks. We see the architecture that comes after.

The Honest Caveat

Visa processes $14 trillion in annual payment volume. They have relationships with every major bank on earth. Their merchant-funded offer model will probably work — within the closed-loop card ecosystem, for issuers who need to differentiate, it's a real upgrade.

Open-loop networks have a harder path. They require brands to trust shared infrastructure. They require consumers to adopt a new mental model. The coalition loyalty programs that tried this before — Plenti, Nectar Italy, early Aeroplan — collapsed under the weight of centralized control and misaligned incentives.

The difference now: decentralized infrastructure that lets brands retain control. Non-expiring value that removes the breakage incentive. API-first architecture that's already visible to AI agents. The demand signal is deafening — nine in ten consumers want it. The infrastructure gap is what's been missing.

Visa is solving for issuers. Someone needs to solve for the customer across all of them.