FollowAgents AI AGENTS DIRECTORY
Home / AI Agent Glossary / Agentic Commerce
Ecosystem & Emerging Terms

Agentic Commerce

Also called: agent-led commerce · agentic checkout · AI shopping agents

Agentic commerce is the emerging practice — and the stack of protocols built for it — in which an AI agent carries out purchases on a user's behalf, from discovering products to checking out and paying, while the person sets the intent and approves the result rather than clicking through each step.

DiscoverAssemble cartUser approvesAgent pays
Agentic commerce shifts the person from clicking through every step to setting intent and approving a cart, while the agent handles discovery, assembly, and payment.

Online shopping today assumes a human at the keyboard: you search, compare tabs, add to cart, and click buy. Agentic commerce is the shift to a model where you state a goal — "reorder my contact lenses," "find a black raincoat under $120 that arrives by Friday" — and an agent does the searching, comparing, cart-building, and payment, coming back to you for approval at the point that matters.

Making that work across many merchants needs shared plumbing, and in 2025-2026 several efforts converged into a rough layered stack: a discovery layer (an agent finds what a business offers, via mcp, webmcp, or structured feeds), an agent-to-agent negotiation layer (a2a-protocol), a commerce/checkout layer for the shopping flow with a merchant (ucp-universal-commerce-protocol from Google and Shopify, or the similarly named Agentic Commerce Protocol from OpenAI and Stripe), and a payment-authorization layer (ap2-agent-payments-protocol) that proves the charge was within what the user allowed.

This is early and unsettled. The protocols overlap, backers differ, and the acronyms collide — "ACP" alone is ambiguous between the commerce protocol and the older Agent Communication Protocol. What is consistent across the serious efforts is the guardrail shape: the user sets spending limits and intent up front, and a specific cart gets explicit approval before money moves.

How it works

A typical flow: the user gives the agent a goal and constraints (budget, delivery date, preferences), often captured as a signed intent. The agent discovers candidate products across merchants through whatever interface each exposes, evaluates them (price, a real checkout quote including shipping and tax, return policy), and assembles a cart at the chosen merchant using a commerce protocol. The user approves that specific cart. A payment layer then authorizes the charge against the user's chosen instrument, with the authorization verifiable by the merchant and payment network. Post-purchase actions — tracking, returns — run back through the same commerce interface.

Example

A household runs low on printer ink. The user has previously told their assistant: "keep the office stocked, under $60 a month, prefer the store-brand cartridges." The agent detects the low-stock signal, checks three retailers through their commerce interfaces, finds the store-brand set at $34 with two-day delivery, assembles the cart, and — because the standing intent is specific and within budget — issues a payment authorization for $34. The user gets a notification, not a checkout page.

How it differs

Agentic commerce vs. a shopping chatbot: a chatbot recommends products and links you out to buy them yourself. Agentic commerce means the agent actually transacts — building the cart at the merchant and authorizing payment — under limits the user set, with a defined approval point, rather than handing the task back to a human at checkout.

Common misconceptions

Often assumed: Agentic commerce means the agent spends your money freely with no oversight.
Actually: The mainstream protocols are built around explicit up-front spending limits and an approval step for the specific cart; an out-of-bounds charge fails verification rather than going through.
Often assumed: There is one agentic-commerce standard.
Actually: As of 2026 there are several overlapping efforts — UCP, the commerce ACP, AP2, and payment extensions — layered rather than unified, and the naming (especially "ACP") is still ambiguous.

FAQ

What is agentic commerce?
The pattern where an AI agent discovers products, builds a cart, checks out, and pays on a user's behalf, with the person setting intent and spending limits and approving the final cart, rather than clicking through checkout themselves.
What protocols are involved in agentic commerce?
Roughly a layered stack: discovery (MCP/WebMCP/feeds), agent-to-agent negotiation (A2A), a commerce/checkout layer (UCP or the commerce ACP), and payment authorization (AP2). They overlap and are still converging.
Is agentic commerce safe — can the agent overspend?
The serious protocols require the user to set limits and intent up front and to approve the specific cart before payment; a charge outside those bounds is rejected at verification. It's still an early, evolving area, though.

Last checked: 2026-08-30

Related terms

AP2 (Agent Payments Protocol)

AP2 (Agent Payments Protocol) is an open protocol for letting an AI agent pay on a user's behalf, structured so that every purchase carries cryptographically signed proof of what the user asked for, what the agent put in the cart, and what may be charged.

UCP (Universal Commerce Protocol)

UCP (Universal Commerce Protocol) is an open standard that lets an AI agent carry out a full shopping journey — discovery, cart, checkout, and post-purchase — against many different retailers through one common, capability-negotiated interface.

WebMCP (Web Model Context Protocol)

WebMCP is a proposed browser API (`navigator.modelContext`) that lets a web page publish JavaScript functions as structured tools a browser-based AI agent can discover and call directly, instead of reading and clicking the page like a person.