Artificial intelligence (AI) agents have arrived, but who is using them across commerce? The software-driven autonomous transaction solutions are adding a new layer to the payments stack, but the marketplace remains split as to whether the momentum will gather around retail or wholesale use cases.
The future being sold by AI providers is one where AI agents will shop, software will negotiate, digital storefronts will sell to machines, and payments will happen without a person tapping a checkout button. That vision is real enough to attract attention. But it may also be too consumer-shaped to explain where the first durable business opportunity sits.
Cloudflare’s new Monetization Gateway, announced at the start of this month, reveals that, for all the attention around “agentic commerce,” a more immediate business opportunity may be less glamorous and more consequential: agentic administration.
Cloudflare said it will let customers charge for web pages, datasets, APIs and Model Context Protocol tools it protects, with payment verification and enforcement settled in stablecoins over the x402 payment protocol, a universal open-source standard created to facilitate agentic payments.
Early participants in the x402 Foundation include Adyen, Amazon Web Services, American Express, Ampersend.ai, Base, Circle, Cloudflare, Coinbase, Fiserv Merchant Solutions, Google, Kakao Pay, Mastercard, Merit Systems, Microsoft, Polygon Labs, PPRO, Shopify, Sierra, Solana Foundation, Stripe, thirdweb and Visa.
For CFOs and merchants, the initiative serves as a sign they can start looking past the agentic shopping near-future where AI agents roam the internet making one-off purchases, and toward a B2B reality where software systems gain the autonomy to pay for the administrative work businesses already consume, approve, meter, reconcile and renew.
See also: Agentic B2B Is Here. Are Your Contracts and Invoices Ready?
Agentic Administration Moves Payments Inside the Workflow for CFOs
For CFOs, the more immediate question is not whether an AI agent can buy sneakers. It is whether software usage, data calls, model inference, content access and machine-to-machine work can become payable events inside normal business operations. If a compliance bot pings a sanctions database, if a procurement workflow buys a supplier-risk score, if an internal AI agent calls a premium API to complete a treasury task, x402 suggests those moments could carry payment instructions by default.
The x402 protocol revives the long-unused HTTP 402 “Payment Required” status code and turns it into a payment protocol for the internet. That matters because the object being sold is not necessarily a product. It may be a request, an answer, a data pull, a support escalation, a chunk of compute, or a software action. The request becomes the transaction.
Traditional commerce assumes a sequence: discovery, cart, checkout, authorization, fulfillment, invoice, reconciliation and support. B2B commerce adds more layers: vendor onboarding, purchase orders, invoice matching, spend approval, tax treatment, contract terms and treasury controls. x402 compresses some of that logic into the request itself. The payment instruction, proof of payment and access decision can all travel in the same web interaction.
The finance question is whether that precision is useful or chaotic. A company might not want 60,000 tiny vendor charges hitting the ledger as separate events. But it may want granular usage data, policy-based spending limits, daily settlement files, vendor-level controls and a clearer link between cost and operational output.
The PYMNTS Intelligence report “Tech on Tech: How the Technology Sector Is Powering Agentic AI Adoption” found a widening agentic readiness gap between tech companies and firms in goods and services, with 75% of tech firms reporting they were extremely familiar with agentic AI, versus 33% of goods firms and 38% of services firms.
See also: Agentic Commerce Shifts B2B Marketplaces From Intermediaries to Infrastructure
Micropayments Are an Agentic Feature, Not The Business Model
The technical breakthrough of both x402 and agentic payments more broadly is that very small payments can become economical. But micropayments alone are not how enterprises usually operate. CFOs do not build controls around infinite streams of isolated pennies. They build around vendors, budgets, approvals, cost centers, policies, audit trails and recurring obligations. That makes the real opportunity look less like pure pay-per-click monetization and more like subscription economics with usage intelligence underneath.
The business internet already solved part of this problem with SaaS. A company does not approve every database query, CRM lookup or cloud compute cycle as a standalone purchase. It approves a vendor relationship, negotiates a pricing model, applies limits, receives usage reporting and reconciles against a billing period.
Cloudflare’s own framing points in that direction. Its gateway is not just a wallet button. It is a rules layer: which route should cost money, which caller should pay, when authentication should be required, how much a variable task should cost, and whether payment enforcement can be managed as infrastructure configuration. That is closer to enterprise billing logic than consumer checkout.
And for merchants and platforms, accepting x402-style payments is not simply a matter of turning on a new rail. It creates new operating questions. Can the payment carry enough metadata to support reconciliation? A finance team will need to know not only that a bot paid 12 cents, but which workflow triggered it, which business unit consumed it, what resource was accessed, what vendor supplied it, what tax treatment applies and whether the transaction belongs against an approved purchase order or contract.
The PYMNTS Intelligence report “How Acquirers Prepare for Agentic Commerce” found that nearly 80% of surveyed acquirers said they are at least somewhat prepared to support seamless omnichannel shopping experiences, a prerequisite for any system in which autonomous agents transact across digital and physical environments.
