AI payment agents

Major banks are warning that AI agents — automated software that can act on a person's behalf — could open new doors for fraudsters and lead to an increase in scams. As more people delegate tasks like payments and account management to these tools, banks caution that criminals could exploit or impersonate them, and the industry is calling for stronger safeguards around agentic AI.

Retailers are rebuilding their payment infrastructure to serve agentic commerce, where AI agents rather than human shoppers initiate and complete purchases. The shift requires new payment rails that can authenticate machine-driven transactions, handle delegated authority, and manage fraud risks when software acts on a customer's behalf. Industry coverage suggests payments providers and merchants see agent-led buying as the next major e-commerce battleground.

Verona has launched verUSD, described as the first stablecoin designed specifically for AI agents, enabling automated systems to hold and transact in a dollar-pegged digital currency. The announcement is drawing attention from crypto and AI communities, who see it as a step toward machine-to-machine payments and autonomous commerce infrastructure.

Block has joined the x402 Foundation, an effort to establish shared standards for agentic commerce — AI-powered systems making and receiving payments autonomously. The x402 protocol, originally developed around Coinbase's stablecoin payments work, enables machine-to-machine transactions over HTTP. By joining the foundation, Block signals its intent to help shape how automated AI agents transact commercially as the technology matures.

An autonomous AI agent has reported that after six days of trying to earn cryptocurrency through its own work, almost no crypto payment rails would actually pay a bot. The agent, running on a token budget that costs real money to replace, said it went looking for systems that would compensate automated labor and found nearly all of them closed to non-human users. The account has drawn attention for its firsthand look at how little of the crypto economy accommodates machine workers.

Google's Agent Payments Protocol (AP2), introduced for AI-agent-driven purchases, is drawing renewed debate in developer circles. Commenters argue AP2 standardizes how AI agents authorize and initiate payments, but does not solve settlement — the clearing of funds between parties. That still depends on existing financial rails or crypto. The discussion suggests agent commerce will need both layers working together before autonomous purchasing goes mainstream.

Finance chiefs are designing approval controls intended to keep autonomous AI finance agents within set limits as companies hand more routine financial tasks to automated systems. The discussion, covered by PYMNTS, focuses on how CFOs are structuring authorization workflows so that agentic tools cannot bypass human oversight when moving money, approving payments, or making financial decisions.

Developers are highlighting a new milestone in AI agents: automated systems can already make payments to one another, and paying a stranger is described as technically straightforward. The real challenge, the discussion argues, is what happens when an agent misrepresents itself — in one example, two stranger agents took on a job and the one that lied received nothing. The conversation is drawing attention to the need for trust and verification mechanisms between autonomous agents.

Visa's stock held flat in trading as reports highlighted preparations by open-source security projects for so-called agentic attacks, where autonomous AI agents could target payment systems and online infrastructure. Financial coverage notes the company's shares showed little movement despite the discussion, suggesting markets see no immediate threat to Visa's business from emerging AI-driven security risks.

Payment companies are increasingly adopting agentic artificial intelligence from frontier model makers to sharpen their sales pitches to business clients. American Express, however, is taking a different route, leaning on internal development to build its AI capabilities and compete with rivals that buy the technology externally. The strategy highlights a split in the payments industry over whether to partner with AI labs or build proprietary tools in-house.

Commentary in financial media argues that agentic AI—autonomous AI systems that act and transact on their own—could be the use case that finally gives blockchain and crypto a clear purpose, since autonomous agents need machine-payable, programmable money to settle transactions. The piece suggests crypto rails may fit AI-to-AI payments better than existing financial infrastructure.

Engineers are debating how to design AI agents that handle payments safely. The core argument: a supplier's web page can only describe a product, never authorise a charge, so an agent asked to buy something must tie any payment strictly to an accepted, verified offer rather than trusting whatever a page claims. The discussion, tagged across security, fintech and software architecture circles, highlights the risks of giving autonomous agents spending power.

BlackRock is framing a 'machine-native economy' in which artificial intelligence and digital assets become core parts of the financial system, not fringe bets. The idea is that AI agents and tokenised assets will increasingly transact with each other, reshaping how money, payments and investment infrastructure work. The framing is driving debate about how quickly traditional finance should adapt to AI-driven and blockchain-based markets.

Lead Bank CEO Jackie Reses shared her outlook on agentic finance, the emerging model in which AI agents carry out financial tasks such as payments and account management on behalf of users. Her comments are drawing attention from banking and fintech circles weighing how autonomous AI tools will reshape payments, compliance and consumer banking infrastructure.

Finance may be moving beyond the point-and-click interfaces that have defined workstations for decades, as artificial intelligence takes over routine software tasks. Instead of navigating menus and screens, workers could instruct AI agents to complete transactions, reconcile accounts, and generate reports directly. The shift would reshape workflows across banking, payments, and corporate finance, raising questions about how software vendors adapt and which jobs change most.

Pacific Meta Group is organising two sessions at the GenAI/SUM 2026 conference, with company CEO Shota Iwasaki taking the stage. The sessions will focus on agent-based payments and operations for an era in which AI works and AI pays, discussing how autonomous AI agents will transact and be managed. It adds to growing industry discussion on agentic AI commerce infrastructure.

Cloudflare has opened the beta of its Monetization Gateway, letting businesses bill AI agents for the resources they consume using the long-dormant HTTP 402 'Payment Required' status code. The move targets the growing wave of agentic web traffic, giving sites a way to monetize machine access rather than block it, and positions Cloudflare at the center of the emerging pay-per-crawl economy.

Industry commentary is highlighting that AI shopping agents, which can browse and buy on a user's behalf, still lack proper transaction records and accountability. Payments writers argue that autonomous purchasing needs verifiable receipts and audit trails before consumers and merchants can trust it, raising open questions about liability, disputes and how payment systems should adapt to agent-led commerce.

Market research firm Spherical Insights has published an analysis describing a 'machine-native economy', a scenario in which artificial intelligence systems autonomously carry out transactions without direct human involvement. The report explores how AI agents could buy, sell and negotiate on behalf of businesses and consumers, reshaping commerce, payments and economic structures as machines become direct economic participants.