As Pay by Bank becomes more visible in the U.S. payments market, one group of consumers is clearly leading the way: those who use digital banks. “Pay by Bank Deep Dive: Digital Bank Users Are Ready to Switch,” a collaboration between PYMNTS Intelligence and Trustly,examines why these users are more open to paying directly from their bank accounts and what needs to happen for that interest to become everyday behavior.
Jeff Bezos famously declared “your margin is my prospect,” encapsulating Amazon’s strategy of leveraging technology, scale, and customer obsession to disrupt incumbents and pass savings onto consumers. In the emerging Prompt Economy, this principle is becoming systemic. Instead of a single company actively seeking out margins to exploit, AI agents, acting on behalf of millions of consumers and businesses, are continuously hunting for margins across all sectors.
PYMNTS Intelligence research demonstrates just how rapidly this shift is unfolding.Nearly 70% of consumers express interest in utilizing AI agents to streamline shopping, with over half desiring an autonomous agent to manage their weekly groceries or even identify thoughtful gifts based on personal connections . PYMNTS estimates that approximately 30 million “Pro” consumers are already relying on generative AI and agentic techniques to handle the majority of 54 everyday tasks, from shopping and bill payments to travel arrangements. These consumers are essentially instructing software to identify and reclaim margins from others.
In this new landscape, the “opportunity” inherent in a margin no longer primarily resides with a platform, but rather with the agent representing the end user. The onus is now on the ecosystem to demonstrate to that agent that any retained margin is justified by tangible value – be it price,convenience,security,or insightful details – or risk having that margin redirected elsewhere.
The clash between human-driven ride-hailing services and autonomous fleets vividly illustrates this dynamic. Uber’s initial success was built on transforming underutilized assets – human labor and privately owned vehicles – into a fluid network. However, this model shifted risk and costs (like driver compensation) onto the drivers themselves. The largest cost within that system is the driver’s time.
Robotaxis represent a fundamental inversion of this logic.A recent analysis revealed that Waymo’s driverless rides in San Francisco average $20, compared to $16 for UberX and $14 for Lyft – a 31% and 41% premium, respectively. Despite the higher cost, demand for Waymo is surging. trip volumes have exploded from just over 12,000 paid rides in August 2023 to over 700,000 monthly by early 2025, accumulating over 10 million paid rides across multiple cities. Surveys indicate that approximately 70% of Waymo riders prefer the driverless experience,with over 40% willing to pay a premium of up to $10 for it .
This demonstrates a clear shift in value perception. Today, the higher cost of a Waymo fare reflects the initial capital and operational expenses of an emerging autonomous network. Though, as fleets scale and technology matures, the elimination of driver costs will unlock critically important efficiencies.The platform’s role is evolving from simply matching riders with drivers to orchestrating demand across a mix of human and autonomous fleets, and increasingly, becoming a direct endpoint for consumer agents and agentic mobility protocols.
Essentially,the driver’s share of the revenue is becoming an opportunity for those who control the autonomous infrastructure,the dispatch algorithms,and the financing behind them.
The payments industry is another arena where this margin-hunting dynamic is playing out. for decades, card economics have been built on interchange fees, breakage, and a complex system of incentives between issuers, networks, acquirers, and merchants. Interchange funds consumer rewards and fraud protection, while merchants view fees as a necessary cost of doing business.
PYMNTS Intelligence research reveals the strong consumer attachment to this model. Roughly 72% of cardholders cite rewards as a key factor in their card selection, with over half strategically choosing cards to maximize those rewards and a quarter rotating cards across categories. Consumers are already actively seeking value, albeit through conventional means.
Open banking and pay-by-bank solutions are often positioned as a merchant-amiable choice to card fees, offering instant account-to-account payments with lower costs and richer data. However, adoption rates remain modest, currently representing only a low single-digit percentage of total consumer payments. Nevertheless, interest is growing, with around 40% of U.S. consumers indicating they would consider pay by bank, particularly younger demographics, for routine debit card purchases.
The key hurdle is replicating the rewards and protections consumers have come to expect from cards. AI agents are poised to change this equation. Instead of merchants unilaterally dictating payment rails, consumers (through their agents) will define the rules, optimizing for net benefits considering rewards, cash flow flexibility, security, and price. This will create a competitive showdown among merchants, issuers, consumers, and networks, as agents calculate the true value of each option.
On the merchant side, retail media and promotional spending represent another significant margin pool at risk. Retailers and platforms have built lucrative advertising networks on top of low-margin product sales, monetizing search placement and digital shelf space using first-party data. analysts predict that retail media will surpass traditional TV ad spend, reaching over $100 billion in global revenue by the end of the decade.
In the Prompt Economy, product discovery will increasingly occur within the agent layer. Rather of browsing retailer websites, consumers will provide agents with specific goals – “new running shoes,” “a four-slice toaster” – along with their preferences and constraints. The agent will then conduct the search, compare prices, check reviews, and vet merchants across multiple platforms .
This shift renders paid placements, co-op promotions, and on-site banners vulnerable. An agent analyzing structured product data, net prices (including fees and promotions), shipping terms, seller reliability, and user preferences will disregard the visual hierarchy imposed by retailers. Any promotional spending that doesn’t demonstrably deliver value will become invisible, transforming the “retail media tax” into an opportunity for agents to secure savings for consumers or demand outcome-based fees from brands and retailers.
this dynamic extends beyond consumer markets. In the B2B realm,AI is disrupting logistics,procurement,trade finance,and treasury management. The margin pools in these areas are even larger, encompassing FX spreads, correspondent banking fees, supply chain financing costs, and inefficiencies in inventory and working capital.
AI-driven planning and optimization are enhancing supply chain predictability and reducing tolerance for inefficiency. Enterprises are leveraging demand forecasting, network optimization, and dynamic routing to minimize inventory and transportation costs.
Agents integrated into ERP and procurement systems can continuously benchmark suppliers based on price, performance, ESG metrics, and risk, automatically reallocating spending when a supplier’s margin is no longer justified by service levels.In trade and treasury, stablecoins and blockchain networks are challenging traditional banking margins, offering near-instant settlement, transparent fees, and programmability.
Banks are responding with tokenized deposits, on-chain cash management, and AI-enhanced trade finance tools, aiming to provide similar speed and programmability with the security and regulatory compliance of traditional banking. This sets the stage for a showdown between non-bank issuers,banks,and corporations,where AI agents become the margin hunters on behalf of businesses.
Consumers and businesses already except certain costs in exchange for convenience, rewards, and efficiency. Consumers pay interest on credit cards, overdraft fees, and delivery charges. Businesses incur FX spreads,slow settlement times,and compliance costs. These are often viewed as the cost of doing business.
Though, the key difference in 2026 and beyond is that AI agents will present consumers and businesses with alternatives, revealing the true, full cost of each option. This openness will transform every hidden or sticky margin into an opportunity for someone else.
Ultimately, the Prompt Economy signals a fundamental shift: business models will be continuously repriced by agents. The Bezos quote remains relevant – but the players have changed. The 2010s saw platforms leveraging data and scale to capture margins. In 2026, agents, intelligent rails, and secure credentials will contest every margin, empowering consumers and businesses to determine who deserves what.
