Legacy Business Models Break

Bezos popularized “your margin is my opportunity” to capture Amazon’s strategy of using technology, scale and obsessive customer focus to attack incumbents’ profit pools and pass much of that surplus back to customers in the form of lower prices and better experiences.

In the Prompt Economy, that instinct becomes systemic: instead of one company hunting margins, AI agents acting for millions of consumers and enterprises hunt margins everywhere, all the time.​

PYMNTS Intelligence’s Prompt Economy work shows how far along this shift already is.

Nearly 70% of consumers say they are interested in using AI agents to simplify shopping tasks; more than half would like an autonomous agent to monitor and do their weekly shopping for them, or look through personal interactions with a friend to identify and purchase a good gift.

PYMNTS estimates around 30 million “Pro” consumers already rely on gen AI or agentic techniques to complete the majority of 54 everyday tasks, including shopping, bill pay and travel. These consumers are effectively telling software to go look for other people’s margins and reclaim them as their own value.​

In this world, the “opportunity” in your margin no longer belongs primarily to a platform; it belongs to the agent that represents the end user. The ecosystem’s job is to convince that agent that any margin it keeps is justified by tangible value — price, convenience, protection, insight — or risk having that margin rerouted to someone else.

Nowhere is “your margin is my opportunity” more visible than in the clash between human‑driven ride‑hail and autonomous fleets. Uber’s original model turned underutilized human labor and privately owned vehicles into a fluid network, with Uber keeping a take rate while pushing asset and labor risk onto drivers. The largest cost in that system is the driver’s time.

Robotaxis invert that logic. A recent analysis of almost 90,000 ride quotes in San Francisco found that Waymo’s driverless rides currently cost an average of $20, versus $16 for UberX and $14 for Lyft. That’s roughly 31% more than Uber and 41% more than Lyft. At peak hours, the gap widens. Using today’s platform math, autonomy looks more expensive, not less.​

And yet, riders are flocking to it. Waymo trip volumes in California and Arizona have exploded from just over 12,000 paid rides in August 2023 to more than 700,000 per month by early 2025, and over 10 million paid rides cumulatively across Phoenix, San Francisco, Los Angeles and Austin. Surveys show that about 70% of riders who have tried Waymo say they prefer the driverless experience, and more than 40% are willing to pay somewhat more for it, with a meaningful minority willing to pay up to 10 dollars extra per ride.​

That’s the margin story in motion.

Today, the “margin” in a Waymo fare reflects high capital and operating costs for an early‑stage autonomous network. Over time, as fleets scale and hardware and operations get cheaper, the absence of a driver unlocks a massive labor pool that software and capital can compete to capture.​

In the Uber era, the platform’s opportunity was the spread between what riders paid and what drivers earned. In the robotaxi era, the driver’s share becomes the opportunity for whoever owns the autonomous supply, the dispatch algorithms and the financing vehicles behind them. The platform’s role shifts from matching riders to drivers, to orchestrating demand across mixed human and robotaxi fleets. And increasingly, exposing itself as an endpoint that consumer agents and agentic mobility protocols can book directly.​

In other words, that driver margin becomes someone else’s fleet opportunity.

Nowhere is this clearer than in payments. For six decades, card economics were built on interchange, breakage and a carefully balanced set of incentives between issuers, networks, acquirers and merchants. Interchange funded rewards and protections that consumers came to expect, while merchants viewed fees as a cost of accessing spending power and conversion.​

PYMNTS Intelligence research shows how tightly consumer behavior has latched onto that model.

Roughly 72% of cardholders say rewards influence their card choice; more than half choose cards strategically to maximize those rewards, and about one in four rotates between cards across categories to extract maximum value. In practice, consumers are already behaving like margin hunters, they are just doing it the old-fashioned way.

Open banking and pay by bank are sold as the merchant’s revenge on card economics. Instant account‑to‑account payments with lower fees and richer data. Yet surveys in the U.S. and Europe suggest early adoption remains modest, as in low single‑digit shares of total consumer payments. Interest, though, is material, with roughly 40% of U.S. consumers saying they would consider pay by bank, especially younger cohorts, for those purchases they consider everyday, debit card purchases.

The catch is rewards and protections. Those same consumers expect pay by bank to behave like the cards merchants want to get rid of, with equivalent rewards, chargeback rights and credit access.​

That is where agents change the equation. Instead of merchants deciding unilaterally which rail to push, consumers (through agents) will set rules. The prompt: “Optimize for my net benefits considering rewards, cash‑flow flexibility, protections and price.” Agents will calculate the all‑in value of each option, setting up a business model showdown across merchants, issuers, consumers and networks.​

On the merchant side, retail media and promotions are the other massive legacy margin pool at risk. Retailers and platforms have built high‑margin ad networks atop low‑margin product sales, monetizing search placement and digital shelf space using first‑party data and closed‑loop attribution. Analysts expect retail media to surpass traditional TV ad spend in some markets, with global revenue heading north of $100 billion dollars by the end of this decade. ​

In the Prompt Economy, discovery will move into the agent layer. Instead of scrolling a retailer’s site or marketplace, consumers will increasingly give an agent a goal “new running shoes,” “fancy toaster with four slots” along with preferences and constraints. The agent will then do the searching, price comparison, review checking and merchant vetting across platforms.​

In that world, paid placements, co‑op promos and on‑site banners suddenly look like margins at risk. An agent that is parsing structured product data, net price (after all fees and promos), shipping and return terms, seller reliability and user preferences has no reason to respect the visual hierarchy on a retailer’s page.

Any promotional spend that does not map to real, measurable value becomes invisible. The retail media “tax” on each sale becomes an opportunity for agents to capture that spend and redirect it as consumer savings, or to demand outcome‑based fees from brands and retailers for incremental, verifiable sales.​

This dynamic is not just about consumers. In B2B, the same logic is already tearing into logistics, procurement, trade finance and treasury. Here, the margin pools are even bigger: FX spreads, correspondent banking fees, credit, supply‑chain financing spreads and the implicit cost of inventory and working‑capital mismanagement.

AI‑driven planning and optimization are making supply chains more predictable and less tolerant of inefficiency. Enterprises are using demand forecasting, network optimization and dynamic routing to squeeze slack out of inventory and transport, reducing the need to pay premiums for last‑minute capacity or buffer stocks.

Agents embedded in ERP and procurement systems can continuously benchmark suppliers on price, performance, ESG metrics and risk, then reallocate spend as soon as a supplier’s “margin” is no longer justified by service levels.​

In trade and treasury, stablecoins and on‑chain networks have already shown how much bank margin is up for grabs. Stablecoin transaction volumes have reached the tens of trillions of dollars annually, with many corporates and platforms attracted by near‑instant settlement, transparent fees and programmability that contrast sharply with the delays and opacity of correspondent banking.

Banks are responding with tokenized deposits, on‑chain cash management and AI‑enhanced trade‑finance tools that promise equivalent speed and programmability with the regulatory comfort and credit relationships corporates value. And setting up a business model showdown among nonbank issuers, banks and corporates where the corporate margin hunter becomes a piece of software.

Underneath all of this is the status quo reality. Consumers and enterprises already pay a lot for the convenience and rewards of legacy models. Consumers pay interest on revolving credit card balances, overdraft and late fees, delivery and service charges on food and grocery orders, and subscription fees for everything from streaming to same‑day delivery. Enterprises pay in the form of FX spreads, slow‑settlement float, insurance and compliance overhead, and suboptimal working‑capital allocation.​

They treat them as the cost of doing business, tradeoffs for efficiency, outcomes and better use of their time. The difference in 2026 and beyond is that agents will show consumers and corporates alternatives, along with the true, full cost of each option. That visibility turns every hidden or sticky margin into someone else’s opportunity.

Taken together, the picture that emerges is not just “legacy business models are under pressure,” but that business models will be continuously repriced by agents.

The Bezos quote is still right — but the protagonists change. The 2010s story was platforms using data and scale to turn other people’s margins into their opportunity. The 2026 story is agents, rails and intelligent credentials turning everyone’s margins into contested territory, with consumers and enterprises in a better position to decide who deserves what.

The post Legacy Business Models Break appeared first on PYMNTS.com.