Agentic Commerce
Agentic commerce is shopping and payment carried out by AI agents on a person's or business's behalf: the agent finds products, compares options and completes a purchase within limits the buyer has authorized, using payment credentials and checkout interfaces built for agents.
AI agents that buy on someone's behalf change who decides, who pays and who answers when something goes wrong. These guides cover the roles, protocols, payment authority, merchant readiness and the operating controls agentic commerce needs.
The full topic archive.
Agentic commerce in B2B procurement: how AI agents can buy for businesses
A procurement agent is only as safe as the policy it carries. Give it a catalog, a budget, a card that enforces both and a record finance can audit.
Who is liable when an AI agent buys the wrong thing? Chargebacks in agentic commerce
The cardholder authorized the agent. Did they authorize this purchase? That question sits between fraud and an ordinary dispute, and the evidence to answer it has to exist before the chargeback arrives.
Agentic commerce glossary: 40 terms explained
One page for the vocabulary of AI agents that shop and pay: protocols, credentials, mandates, authentication and merchant operations, each defined in a sentence or two.
Agentic commerce in the UAE and Saudi Arabia: what merchants and banks should prepare
Agent-ready checkout in the Gulf means more than accepting international cards. It means local schemes, instant account payments, consent rules and bilingual data agents can read.
Agentic vs conversational vs headless commerce: what is the difference?
The three terms answer different questions: where the customer talks, how the store is built and who is allowed to decide. They often appear together, but they are not the same thing.
How to prepare an online store for AI shopping agents: a merchant checklist
Agents read your product data, not your homepage. Get the feed accurate, decide where checkout happens, accept scoped payment credentials and keep order updates flowing.
Can AI agents pay with stablecoins? How agent stablecoin payments work
Stablecoins suit payments that are small, frequent and machine to machine. They come without card-style disputes, so limits and records have to be designed in.
What is the Agentic Commerce Protocol (ACP)? How AI agents check out
ACP gives an agent a checkout session it can call and a payment token capped to one purchase. The seller stays merchant of record and keeps its own processor.
What is agentic commerce? A plain-English guide to AI agents that shop and pay
An AI agent that can buy needs three things a chatbot does not: authority from the buyer, a checkout it can call and a payment credential scoped to the purchase.
What is AP2? Google's Agent Payments Protocol explained
AP2 does not move money. It produces signed proof of what a user let an agent buy, so the merchant, the credential provider and the network can each check it before anything is charged.
What is Mastercard Agent Pay? Agentic tokens explained
Mastercard's answer to agent payments is a registered agent holding a token, not a bot holding a card number. Here is how agentic tokens, agent verification and merchant acceptance fit together.
What is Visa Intelligent Commerce? Visa's agentic payments program explained
Visa does not give an AI agent your card number. It gives the agent its own token, asks you to approve each instruction with a passkey, and checks the purchase against what you approved.
What is x402? HTTP 402 payments for AI agents explained
x402 turns a status code that sat unused for decades into a payment step: the server answers 402 with a price, the agent signs a payment and retries, and the resource comes back with a receipt.
When Commerce Needs an Agent, and When a Workflow Is Enough
Choosing among changing offers may need judgment. Enforcing a purchase limit and recording the order still need dependable controls. Decide where autonomy earns its cost.
Two agents need one durable purchase intention
Two workers can create two orders from one approved task. Give the purchase a durable identity and keep recovered outcomes away from fresh submission.
Agentic Commerce From Purchase to Refund: Follow One Order
An agent can find a product and start checkout, but each later step needs its own authority and evidence. Follow one purchase through delivery and a partial return.
An Agent Can Find the Product. May It Buy It?
A request to compare monitors does not authorize a purchase. The useful boundary is the exact instruction that still holds when the agent is ready to act.
Review an agentic commerce pilot before scaling
An 87.5% completion rate needs a traceable denominator. A worked pilot scorecard separates requests, attempts, authorized orders and unresolved outcomes.
What may a shopping agent decide after the purchase?
An agent may be able to read an order without permission to accept store credit or a fee. Evaluate the proposed action against the current customer instruction.
MCP, A2A, ACP, AP2 and UCP: Check the Contract Behind the Name
A merchant and an agent can support the same protocol name while disagreeing about its version, payment handler or supported operation. Resolve that contract before accepting a purchase.
Test a UCP checkout from offer to confirmed order
An agent can create a checkout without completing a purchase. Test the buyer handoff, refreshed state and merchant outcome for one pinned protocol version.
Agentic Commerce: Which Agent Does What?
A shopping assistant can recommend a purchase without having permission to make it. Map the responsibilities, tools and evidence across one agent-assisted purchase.
Agentic Commerce: Payment Product Readiness
An agent purchase needs an enforceable spending instruction, a reliable order and payment lifecycle, and someone who owns the exceptions. Start with a bounded buying task and test the failure paths.
Agentic Payments Need an Identity Control Plane
Agent-led checkout will not scale on payment authorization alone. Banks, merchants and platforms need evidence about who the agent is, who it represents and what it was allowed to do.
KYA Makes Agent Payments An Identity Control Plane
Know-Your-Agent will matter only if payment operators can prove who controls the agent, what it is allowed to do, how it was certified, and which transaction evidence survives exceptions.
Mastercard Agent Connect Makes Merchant Control The Product
Agentic commerce will not scale because AI agents can shop. It will scale when merchants can control participation, product data, pricing, fulfilment, payment authorization and exception evidence.
Solana Pay Shows Agent Payments Need Wallet Approval Controls
Solana Foundation's pay CLI is a repo-radar signal for agentic payments: the useful product boundary is not automatic payment, but local wallet approval, policy checks, receipts, and denial before signing.
Agent Payment Guard Shows x402 Needs Pre-Payment Risk Gates
Agentic payments do not become safe because the payment rail works. They become safe when every agent payment has an approved mandate, bounded amount, trusted counterparty, and a pre-signing risk gate that can stop the transaction.
The UK Financial Services AI Plan Is a Delivery Governance Test
The UK financial-services AI plan is not just policy. For banks and fintechs, it is a programme governance test across models, vendors, skills, resilience, and agentic payments.
Cross River and Stripe Show Why Agentic Cards Need a Mandate Ledger
A single-use virtual card can protect credentials. It cannot, by itself, prove that an agent stayed within the user's mandate.
Forter Agents Show AI Risk Work Is Becoming Operational
Forter's agent launch and today's repo radar point to the same pattern: AI is moving from generic assistants into bounded workflows with data access, controls, and operating accountability.
Agentic Commerce: What Visa and Mastercard Are Really Building
A shopping agent that compares, selects, and pays under authority you set is a new economic actor. Visa, Mastercard, OpenAI, and Stripe are racing to build the trust layer that lets merchants and issuers accept it.
Agentic Payments Operations: What Works, What Is Theatre
Agentic AI can help payments operations when the task is bounded, observable and reversible. It becomes theatre when teams let agents improvise inside money movement.
Common questions about Agentic Commerce.
What is agentic commerce?
Commerce in which an AI agent acts for a buyer: discovering products, building a cart and paying within the authority the buyer has granted. It differs from conversational commerce, where a person makes each decision in a chat, because the agent can complete steps on its own within agreed limits.
How do AI agents pay for things?
Through payment credentials that are scoped or issued for agent use and checkout interfaces agents can call. Stripe, for example, describes shared payment tokens that give a seller scoped access to a customer's payment method for an agent-initiated purchase.
Which protocols are used in agentic commerce?
Several open specifications are emerging, including the Agentic Commerce Protocol for agent checkout, the Universal Commerce Protocol for interoperable checkout, orders and payment-token exchange, the Agent Payments Protocol for payment mandates, and MCP and A2A for connecting agents to tools and to each other. Check which version a partner actually supports.