Agentspay

Explainer

ChatGPT Instant Checkout: What Happened to OpenAI Instant Checkout, and What Replaced It

Almost every guide to ChatGPT Instant Checkout still reads like it was written the week it launched, walking merchants through how to apply and what the fee will be. OpenAI changed course in March 2026. Here is the accurate version: what Instant Checkout was, what the numbers actually looked like, what replaced it, and which parts of the stack are still very much alive.

Agent Payments Console

Pick an agent

Payment intent

intent:

Policy evaluation

Human approval required

This spend is over your approval threshold. Approve it to issue a scoped card, or deny it.

Scoped virtual card issued

Agentspay

single-use

Wallet budget

spent of

Audit trail

In short

ChatGPT Instant Checkout was a feature that let a shopper complete a purchase inside a ChatGPT conversation, without visiting the merchant’s website. OpenAI launched it on September 29, 2025 with Etsy, then extended it to Shopify merchants, PayPal and Walmart. In early March 2026 OpenAI phased it out as a standalone in-answer checkout, telling reporters it was "evolving how we approach commerce in ChatGPT" and that "Instant Checkout is moving to Apps, where purchases can happen more seamlessly." The trigger was performance: Walmart EVP of product and design Daniel Danker said in-chat purchases converted at roughly one third the rate of sending the same shopper to walmart.com, and Forrester principal analyst Emily Pfeiffer counted only about 30 participating Shopify merchants as of February 2026. The Agentic Commerce Protocol underneath it was not retired. OpenAI still documents ACP as live infrastructure, and checkout now happens inside merchant-branded apps in ChatGPT such as Target, Instacart, DoorDash and Walmart’s Sparky.

What is OpenAI Instant Checkout?

Instant Checkout was OpenAI’s attempt to close the loop between product discovery and purchase inside ChatGPT. A shopper asked for something, ChatGPT returned product results, and one of them carried a Buy button. Tapping it opened a checkout surface in the conversation with live shipping options, tax and a total pulled from the merchant’s own systems. The shopper confirmed, the order landed in the merchant’s back office, and the merchant stayed the seller of record the whole time. That last detail mattered more than it sounds: OpenAI never took possession of the sale, never became the retailer, and never replaced anyone’s payment processor. It rendered a checkout and passed a scoped, one-time payment credential to the merchant, who charged it through their existing provider exactly as they would any other card transaction.

When did ChatGPT Instant Checkout launch?

OpenAI shipped Instant Checkout on September 29, 2025, alongside Stripe, and released the Agentic Commerce Protocol as an open standard on the same day. Etsy sellers were live first. Shopify merchants followed, PayPal onboarded its own base of small businesses through an ACP server, and in November Walmart made roughly 200,000 products available through the feature. On paper that is a formidable opening roster. In practice the distribution was far narrower than the headline partner names suggested, and the gap between those two facts is the whole story of what happened next.

Did OpenAI abandon Instant Checkout?

Not abandoned, but genuinely deprioritized, and the distinction matters if you are deciding what to build. In early March 2026 an OpenAI spokesperson said the company was "evolving how we approach commerce in ChatGPT to better meet merchants and users where they are," and added that OpenAI was "prioritizing making ChatGPT search and product discovery great, with ACP serving as the infrastructure that connects users to merchants across the full shopping journey. Instant Checkout is moving to Apps, where purchases can happen more seamlessly." Read that carefully. The in-answer Buy button is what went away. The protocol underneath it was explicitly reaffirmed, and buying inside ChatGPT did not stop, it relocated. On March 24, 2026 OpenAI followed up with a rebuilt shopping experience in ChatGPT weighted toward finding and comparing products rather than transacting on the spot.

Why did Instant Checkout fail?

Two numbers explain it better than any amount of strategy commentary. The first is conversion. Daniel Danker, Walmart’s EVP of product and design, said purchases completed inside ChatGPT converted at about one third the rate of simply clicking through to walmart.com and buying there, and he described the in-chat experience as "unsatisfying". Losing two thirds of your conversion in exchange for a slightly shorter path is not a trade a large retailer makes twice. The second is adoption. Forrester principal analyst Emily Pfeiffer counted roughly 30 Shopify merchants actually participating as of February 2026, out of a merchant base in the millions. Announced availability and live availability turned out to be very different things. Shopify president Harley Finkelstein put his finger on the underlying reason when he pointed out that transaction integrity is not only about payments, it is "checkout itself, subscriptions, inventory, shipping taxes, all the different merchandising options." A generic checkout surface rendered by a third party cannot carry all of that, and merchants noticed.

What replaced Instant Checkout in ChatGPT?

Apps. Instead of OpenAI rendering a universal checkout, merchants now bring their own branded experience into the conversation, and the purchase completes in the merchant’s own environment where their merchandising, loyalty, subscriptions and returns policy all still work. Target, Instacart, DoorDash and The Knot are among the retailers that launched apps in ChatGPT. Walmart is embedding Sparky, its own assistant: a shopper logs into Walmart, the cart syncs between the two surfaces, and checkout happens on Walmart’s stack. The industry shorthand that emerged for this is discover in AI, buy on site. It is a smaller claim than the original pitch, and it is a much better fit for how large retailers actually protect conversion.

Is the Agentic Commerce Protocol dead?

No, and this is the part most coverage gets wrong by omission. ACP is the open standard OpenAI co-developed with Stripe, released under Apache 2.0, and it was named directly in OpenAI’s own March statement as the infrastructure connecting users to merchants. As of August 2026 OpenAI’s commerce documentation still describes ACP as current with no deprecation notice: merchants publish a regularly refreshed product feed as CSV or JSON, expose endpoints that manage a checkout session, and receive a one-time delegated payment request carrying a maximum chargeable amount and an expiry. The merchant validates, prices, taxes, accepts or declines, charges through their own PSP and fulfills. Instant Checkout was one product built on ACP. Losing the product did not remove the plumbing, and the plumbing is what a merchant integrates against.

What did ChatGPT Instant Checkout cost merchants?

This deserves more care than it usually gets, because the number everybody repeats was never confirmed at the source. A 4% fee on completed orders, charged by OpenAI on top of ordinary card processing, was widely reported, and reported specifically in the context of Shopify merchants. Sam Altman has separately referred publicly to a rate closer to 2%. OpenAI never published a rate card. There were no setup fees, no monthly platform fee, and nothing extra charged to the shopper, so the honest summary is that the take rate sat somewhere in a 2% to 4% band on top of processing, varied by channel, and applied only to completed sales. If you are modeling the economics of an agentic sales channel today, use that range as a planning assumption rather than a quoted price, and expect app-based routes to price differently since the checkout is yours.

What should merchants do now?

Stop treating in-answer checkout as the destination and start treating structured product data as the asset. Everything in this space, ACP, the Universal Commerce Protocol that Google launched on January 11, 2026, and whatever arrives next, consumes the same input: a clean, current, machine-readable feed with accurate identifiers, prices, inventory, shipping and returns terms. Merchants that already have that can turn on a new agentic channel in days. Merchants that do not will spend a quarter on data cleanup regardless of which protocol wins. Beyond the feed, the practical checklist is short. Make sure your site renders product facts server side so agents can read them. Decide whether an app in ChatGPT is worth the build for your category. And fix attribution now, because orders arriving from agent surfaces show up in analytics with no session history and quietly break every acquisition model you have. If you are on Shopify, most of that checklist is handled by the platform and the remaining work is data quality and fees, which we cover in the Shopify agentic commerce guide.

What this means if your company buys, rather than sells

Here is the shift almost nobody covered. The Instant Checkout retreat was reported as a merchant story, and for the buy side it changed nothing that matters. Your employees and your shopping agents are still purchasing through AI surfaces. The purchase simply moved from a checkout OpenAI rendered into a merchant app, which from a finance perspective is strictly worse, because it is even further from anything your systems observe. The card still gets charged. The subscription still renews. The line item still lands on a statement three weeks later with a merchant name nobody recognizes. Whether the checkout screen lives in ChatGPT, in a Target app inside ChatGPT, or on target.com, the control question is unchanged: which agent was allowed to spend what, who approved the exception, and where is the record.

Why no checkout protocol solves the budget problem

Every standard in this space caps a single transaction and none of them tracks a total. ACP scopes one delegated payment credential to one order with a maximum amount and a short expiry. Google’s AP2 binds a closed mandate to the open mandate it derives from, so one purchase cannot exceed what was authorized. x402 lets a seller cap the price of one call. All of that is genuinely good engineering, and all of it is per-transaction. Nothing in any of these specifications knows that this agent already spent four thousand dollars this month across six merchants and two rails. Twenty individually authorized purchases still add up to a number nobody approved, which is precisely the failure mode a signed mandate cannot catch.

Where a control plane fits

Agentspay sits above the rails rather than inside any one of them. Every agent gets its own wallet with a hard limit, so it is structurally incapable of spending past its budget no matter which surface it buys through. Purchases above a threshold you set pause for human approval before money moves rather than after. Scoped virtual cards restrict a given agent to named merchants and amounts. And every decision, the policy that allowed it, the agent that requested it and the human who approved it, lands in one immutable audit trail that spans card spend, stablecoin settlement and metered API calls together. That is deliberately the layer that does not care whether OpenAI ships an in-answer checkout, an app, or something else entirely next year.

What to watch next

Three things worth tracking. First, whether app-based checkout converts materially better than in-answer checkout did, because if it does not, the discover-in-AI-buy-on-site pattern hardens into the permanent shape of agentic commerce and the protocols become a product-data story rather than a payments story. Second, whether ACP and UCP converge or fragment, since merchants will not maintain two feed formats and two endpoint sets indefinitely. Third, and most relevant to anyone deploying agents rather than selling to them, whether any of these standards ever grows a notion of an organizational budget. So far none has, none is scoped to, and the companies running agent fleets are solving it a layer up. We keep this page current as that changes.

Whatever standard moves the money, Agentspay is the rail-neutral control plane that keeps it governed. See how it works and the control surfaces that enforce policy, approvals, and audit on every transaction.

The timeline

ChatGPT Instant Checkout, launch to phase-out

Dates matter here because most material online still describes the September 2025 version of this feature as current.

Date What happened
September 29, 2025 OpenAI launches Instant Checkout with Stripe and publishes the Agentic Commerce Protocol under Apache 2.0. Etsy sellers live first.
Late 2025 Shopify merchants and PayPal’s small-business base join through ACP servers.
November 2025 Walmart makes roughly 200,000 products available through Instant Checkout.
February 2026 Forrester counts about 30 Shopify merchants actually participating.
Early March 2026 OpenAI deprioritizes Instant Checkout as a standalone feature. Checkout moves to Apps. ACP reaffirmed as infrastructure.
March 24, 2026 OpenAI ships a rebuilt ChatGPT shopping experience focused on discovery and comparison.

Before and after

In-answer checkout versus app-based checkout

What actually changed for a merchant, and what did not change at all.

Instant Checkout (2025 to early 2026) Apps in ChatGPT (now)
Where the purchase completes A checkout surface OpenAI renders inside the chat The merchant’s own branded app experience
Who is seller of record The merchant The merchant
Who processes the card The merchant’s existing PSP The merchant’s existing PSP
Merchandising, loyalty, subscriptions Largely unavailable The merchant’s own, fully intact
Take rate to OpenAI Reported in a 2% to 4% band, never published Set by the app arrangement, not a flat platform fee
Underlying standard Agentic Commerce Protocol Agentic Commerce Protocol, still current
Visibility for the buyer’s finance team None None

Honest scope

What the checkout standards control, and what they do not

Each of these caps one transaction. None of them keeps a running total across agents, merchants or rails.

ACP AP2 x402 A control plane
Caps a single purchase Yes, delegated credential with a max amount Yes, closed mandate bound to an open one Yes, seller caps the call price Yes
Tracks cumulative spend per agent No No No Yes
Routes exceptions to a human No No No Yes
Works across card, stablecoin and API spend No, card rails Cards first, others planned No, stablecoins Yes
Single audit trail across all of it No No No Yes

Frequently asked

Questions people ask about ChatGPT Instant Checkout

Is ChatGPT Instant Checkout still available?

Not as the standalone in-answer Buy button it launched with. OpenAI deprioritized that feature in early March 2026 and moved purchasing into merchant-branded apps inside ChatGPT. You can still buy through ChatGPT, but the checkout now runs in the merchant’s own app experience rather than in a surface OpenAI renders.

How does ChatGPT Instant Checkout work?

A shopper asked ChatGPT for a product, saw results, and tapped Buy on one of them. ChatGPT called the merchant’s endpoints to build a checkout session with live shipping and tax, then passed a one-time delegated payment credential with a maximum amount and a short expiry. The merchant charged it through their own processor and fulfilled the order as seller of record.

What was the ChatGPT Instant Checkout fee?

OpenAI never published a rate card. A 4% fee on completed orders was widely reported, specifically for Shopify merchants, and Sam Altman has publicly referred to a rate closer to 2%. There were no setup or monthly fees and nothing extra charged to the shopper. Treat 2% to 4% as a planning range rather than a quoted price.

Why did Walmart stop using ChatGPT Instant Checkout?

Conversion. Daniel Danker, Walmart’s EVP of product and design, said purchases completed inside ChatGPT converted at roughly one third the rate of clicking through to walmart.com, and called the experience unsatisfying. Walmart is instead embedding its own Sparky assistant into ChatGPT so shoppers log in, sync carts and check out on Walmart’s own stack.

Is the Agentic Commerce Protocol still supported?

Yes. ACP is Apache 2.0, co-developed by OpenAI and Stripe, and OpenAI named it directly as the infrastructure it is continuing to build on. As of August 2026 OpenAI’s commerce documentation describes ACP as current with no deprecation notice. Instant Checkout was a product built on ACP, not ACP itself.

Should merchants still integrate ACP?

If you sell in categories people research through AI assistants, yes, but integrate for the product feed and the order endpoints rather than for one checkout surface. The feed is the reusable asset. It is the same input ACP, Google’s Universal Commerce Protocol and any future standard all consume, so the work carries forward regardless of which one wins.

What is the difference between Instant Checkout and Universal Commerce Protocol?

Instant Checkout was a consumer feature in ChatGPT. UCP is a competing open standard from Google, launched January 11, 2026, that orchestrates the purchase lifecycle across Google’s AI surfaces. The closer comparison to UCP is ACP, the protocol underneath Instant Checkout, not the feature itself.

Does agentic checkout give my finance team any spending controls?

No. Every checkout standard caps one transaction and none tracks a cumulative total across agents, merchants or rails. If your company runs agents that buy things, the per-agent budget, the approval threshold and the audit trail have to come from a control plane above the rails, because no checkout protocol is scoped to provide them.

Keep reading

More explainers

ServiceNow AI Control Tower

ServiceNow AI Control Tower

ServiceNow AI Control Tower is the most complete agent inventory and risk console a large US enterprise can buy, and it now reaches across AWS, Google Cloud and Azure. We read the schema ServiceNow ships to developers to answer the one question the rollout meeting always ends on: can it stop an agent from spending money? It cannot, and the reason is written into the data model.

Read

Gemini Enterprise

Gemini Enterprise

Google did something in August 2026 that the other agent platforms have not done: it shipped a hard monthly spend cap that genuinely stops usage instead of emailing you about it. That deserves credit, and it also moves the interesting question one step along. A cap that stops something is only as useful as the thing it is scoped to, so we went and measured what Google can actually point that cap at, in the API model Google publishes for anyone to read.

Read

Salesforce Agentforce

Salesforce Agentforce

Agentforce is the largest agent platform any US enterprise is likely to already own, and it moved to consumption billing, which means the meter now runs on what your agents do rather than on how many seats you bought. That raises a finance question the rollout deck rarely answers: when an Agentforce agent is loose in production, what actually stops it spending. We went and measured the answer in Salesforce own published object model rather than guessing at it.

Read

AWS AgentCore

AWS AgentCore

Amazon shipped the missing piece in August 2026. Bedrock AgentCore Payments went generally available, and it is a real payments product: an agent can now hold a wallet, meet an HTTP 402, pay, and carry on reasoning without a human in the loop. So the question a platform lead has to answer stopped being whether AWS gives agents money and became a narrower, more awkward one: how much of a spend policy did AWS actually ship? We went and measured it, property by property, in the API model AWS publishes.

Read

Microsoft Agent 365

Microsoft Agent 365

Microsoft shipped a control plane for AI agents, and it is a good one. It gives every agent an identity, a registry entry, an owner, a sponsor and a Conditional Access policy. Then somebody in finance asks the obvious follow-up question: fine, but what stops the agent from spending money? This page answers what Agent 365 costs, what it governs, and what we measured when we went looking for a dollar amount anywhere in Microsoft's agent governance surface.

Read

QuickBooks MCP Server

QuickBooks MCP server

Connecting an accounting system to an AI assistant is now a ten minute job. Deciding what that assistant is allowed to do once it is connected is the part nobody writes about, and it is the part your controller will ask about first. This page compares what the official QuickBooks, NetSuite and Xero MCP servers actually hand a model, measured rather than summarized from marketing pages.

Read

Payment MCP Servers

payment MCP servers

Every large payment company shipped an MCP server in the last eighteen months, and almost every write-up of them is a setup tutorial. The setup is the easy part. The question worth answering before you connect one to a production account is narrower and much less comfortable: what, exactly, can the model on the other end of that connection do to your money?

Read

PayPal Agentic Commerce

PayPal Agentic Commerce

PayPal made a bet that most merchants would rather not implement a commerce protocol at all. Where Stripe and OpenAI shipped a spec for you to build against, PayPal shipped two products that sit on top of the checkout you already have, and then bought a company to make the catalog half work. That choice is the whole story: it explains why Agent Ready needs almost no engineering from you, why there is nothing for an agent to discover about your store on the open web, and why the thing PayPal will not do for you is the thing that gets expensive later.

Read

Shopify Agentic Commerce

Shopify Agentic Commerce

Shopify switched agentic commerce on by default, so your store is probably already selling to AI assistants whether or not anyone on your team configured it. Instead of restating the announcement, we checked something you can check too: on September 2, 2026 we requested the machine-readable capability file that Shopify publishes for real storefronts, on fourteen well-known US brand domains, and read what it exposes to an agent. Eleven answered correctly. The three that did not share one trait, and it is quietly costing them agent traffic.

Read

Web Bot Auth

Web Bot Auth

Web Bot Auth is the reason your agent either gets served or gets throttled with the scrapers. Almost everything written about it repeats the same architecture diagram, so we did something different: on September 1, 2026 we fetched the published key directories of more than twenty major AI operators and infrastructure vendors to see who is genuinely signing their traffic. Four were. The results are in the first table.

Read

Tempo Blockchain

the Tempo blockchain

Tempo is the payments chain Stripe and Paradigm built, and it shipped with a protocol that lets software pay for things on its own. It settles machine payments in under a second. It has nothing at all to say about whether your agent should have paid.

Read

AI Agent Governance

AI agent governance

Every agentic AI governance framework published so far governs the same four things: identity, tools, data and prompts. Not one of them carries a budget. Here is what the real frameworks say, which guardrails actually bind at runtime, and what to do about the last mile none of them reach.

Read

A2A Protocol

A2A Protocol

Most explanations of the A2A protocol stop at the sentence that agents can now talk to each other. That was true in April 2025 and it is no longer the interesting part. A2A shipped version 1.0 in April 2026 under Linux Foundation governance, it runs in production inside Azure AI Foundry and Amazon Bedrock AgentCore, and the questions engineers actually get stuck on are narrower: what an Agent Card commits you to, when to reach for MCP instead, and what happens the first time one of your agents has to pay another one for the work. That last question has a specific answer, and it is not in the core spec.

Read

Mastercard Agent Pay

Mastercard Agent Pay

Nearly every article about Mastercard Agent Pay is a retelling of one press release from April 2025, the one where Mastercard said AI agents would be able to shop with Agentic Tokens and named Microsoft as the first platform. That was sixteen months ago, and four more things have shipped since. Reading only the launch coverage leaves you with roughly a quarter of the picture, and the missing three quarters are the parts that decide whether you can actually put this into production.

Read

Visa Intelligent Commerce

Visa Intelligent Commerce

Almost everything written about Visa Intelligent Commerce is a retelling of the April 2025 announcement, when Visa said AI agents would be able to pay with a Visa credential. Three more things have shipped since, including an open agent-identity protocol built with Cloudflare that most coverage does not mention at all. This page is the current version, checked against Visa’s own developer documentation and newsroom in August 2026.

Read

Stripe agentic commerce

Stripe agentic commerce

Most writing about Stripe and agentic commerce is still a retelling of the September 2025 launch week, when Stripe and OpenAI shipped Instant Checkout and published the Agentic Commerce Protocol together. Stripe has built a good deal more since then, and some of it points in a direction the launch coverage never anticipated. This page is the current version, checked against Stripe’s own documentation in August 2026.

Read

Google AP2

Google AP2

Most guides to Google AP2 still describe an Intent Mandate and a Cart Mandate, because most of them are rewrites of the September 2025 launch post. The specification moved. Here is what the Agent Payments Protocol actually defines today, and the one question it deliberately does not answer.

Read

Human in the loop AI

Human in the Loop AI

Every guide to human in the loop AI describes the same shape: the agent pauses, a person decides, the agent continues. The shape is right. What almost none of them ask is a harder question, which is where the pause is enforced, because a pause written into the agent's own code is a pause the agent is trusted to honor.

Read

AI agent cost

AI Agent Cost

Every cost guide for AI agents answers the same two questions: what does it cost to build, and what does it cost to run. Both are answerable, and both are on somebody's invoice. The third question is the one that ends up in a variance report, because the agent also spends your money, and nobody sends you a bill for that.

Read

Agentic checkout

Agentic Checkout

Nearly every guide to agentic checkout is written for the merchant who wants to receive these orders. Far fewer are written for the company whose agents are placing them, which is odd, because agentic checkout quietly removes the one screen where spending used to get a second look.

Read

API monetization

API Monetization

Most guides to API monetization argue about which pricing model wins. The harder question in 2026 is who is calling. An API priced for a signed-up developer with a key behaves very differently when the caller is an agent that showed up once, wants one record, and has no account.

Read

x402 protocol

x402 Protocol

x402 took the one HTTP status code the web never used and turned it into a payment rail machines can drive. The protocol is elegant and genuinely small. The part it deliberately leaves to you is the budget.

Read

AI procurement agents

AI Procurement Agents

Every major procurement suite shipped agents during 2026. Almost none of them answer the question your controller will ask first, which is what happens when the agent is wrong about a purchase and the money has already moved.

Read

Agentic payments

Agentic Payments

Agentic payments move money with no human at the checkout. The rails to do it all shipped during 2026. The part most teams have not solved is deciding, before the money moves, whether the agent was allowed to spend it.

Read

AI agent monetization

AI Agent Monetization

Every AI agent company is rewriting its price list. The models that survive are metered. The ones that quietly fail are the ones where nobody measured what a single task costs to serve.

Read

Agent payment platforms

AI Agent Payment Platforms

Five different kinds of product now call themselves an AI agent payment platform, and they solve five different problems. Picking the wrong category is the expensive mistake, not picking the wrong vendor inside a category.

Read

Universal Commerce Protocol

the Universal Commerce Protocol (UCP)

Google and Shopify shipped UCP as an open standard so an AI agent can check out at any merchant that supports it. Here is what the specification actually defines, where it is live for US buyers, and the one thing it deliberately leaves to you.

Read

MCP Payments

MCP Payments

MCP payments are how an AI agent discovers a payment tool and calls it to move money. The catch: the Model Context Protocol carries the tool call, not the spending decision, so nothing in the stack asks whether the purchase should have happened.

Read

Visa Intelligent Commerce vs Mastercard Agent Pay

Visa Intelligent Commerce vs Mastercard Agent Pay

Visa Intelligent Commerce and Mastercard Agent Pay are the two big card networks racing to let AI agents pay. They take different routes to the same idea, and neither one decides whether a given purchase should have happened.

Read

Agentic Commerce Protocol

the Agentic Commerce Protocol

ACP is the open standard behind agentic checkout in ChatGPT. It tells a merchant how to sell to an AI agent. It says nothing about whether your agent should have made the purchase.

Read

AP2 vs ACP vs x402

AP2 vs ACP vs x402

AP2, ACP, and x402 are the three standards shaping how AI agents pay. They solve different layers of the problem, and most real systems will touch more than one.

Read

Machine payments protocol

Machine payments protocol

As software starts paying software, machine payments protocols define how value moves without a human at the keyboard. The harder question is how to keep that spending governed.

Read

Know Your Agent (KYA)

Know Your Agent

KYA, or Know Your Agent, extends the idea of customer due diligence to autonomous software. When an agent spends, you need to know which agent, on whose authority, and under what limits.

Read

Keep agent spending governed

Add policy, hard limits, human approval, and an immutable audit trail across any protocol or rail. Start in the sandbox today.

Never moves money without policy