Yes, it’s now possible to shop directly within the AI chat: With “Instant Checkout,” OpenAI launched in-app purchasing in ChatGPT on September 29, 2025—and switched to a model in early March 2026 where the AI finds products, but the purchase is completed with the retailer. For online store owners, this means the store remains the location of the transaction—provided that its product data is readable by agents.
At a Glance
- September 29, 2025: Launch of “Instant Checkout” in ChatGPT, based on the open Agentic Commerce Protocol (ACP) developed with Stripe.
- Early March 2026: Transition to “merchant-controlled checkout”—discovery by the AI, completion in the merchant’s store.
- Three competing approaches: ACP (OpenAI/Stripe), UCP (Google), and Perplexity’s own shopping path—no clear winner has emerged yet.
- Transaction volume via AI agents in the DACH region has been low so far (as of August 2026)—robust preparation: machine-readable product data.
Can you shop directly in ChatGPT?
Yes—since September 29, 2025. “Instant Checkout” made it possible to make purchases directly within a ChatGPT conversation, initially with select U.S. retailers, processed via the infrastructure developed with Stripe. A little over a year later, the picture is more nuanced: OpenAI has redesigned the model, and in the DACH region, direct in-chat purchasing has not yet become a mass phenomenon (as of August 2026). However, the direction is clear—all major AI providers are working on turning product recommendations into transactions.
What is the Agentic Commerce Protocol (ACP)?
The ACP is an open specification from OpenAI and Stripe: It describes how an AI agent reads product data, submits an order, and authorizes payment—without the retailer losing control over its product selection, pricing, and fulfillment. It continues to exist as an open proposal even after the strategy shift; at the same time, Google (UCP) and Perplexity are developing their own shopping solutions. Anyone relying on just one protocol today is taking a gamble—but those who keep their data clean are prepared for any of them.
What does the March 2026 strategy shift mean?
It shifts the balance of power in favor of merchants. With “merchant-controlled checkout,” AI handles discovery, while the purchase is completed in the merchant’s checkout—using the merchant’s payment methods, shipping logic, and customer relationship. At the same time, what the AI sees before the click matters: product name, price, availability, and shipping terms must be machine-readable and up to date—otherwise, the agent will recommend the competitor with the better data.
How do I prepare my store for AI shopping agents?
Independent of the platform—via the data source: The same product data powers Google Shopping and price comparison sites today, and will power shopping agents tomorrow.
| Preparation Steps | Specific Action |
|---|---|
| Feed Quality | Complete titles, descriptions, categories, and images in the merchant feed. |
| Identifiers | Maintain GTIN/EAN and MPN for each item—agents match products using unique identifiers. |
| Price & Availability | Markup, feed, and page must match and be up to date. |
| Structured Data | Product JSON-LD on every product page, including shipping and return information. |
| Stable URLs | Permanent product URLs with redirect history. |
| Bot Policy | Check whether CDN, WAF, or robots.txt rules block AI agents—those that block them remain invisible to us. |
Is it worth the effort right now?
Yes—but for the right reason. Investing in agent revenue, which is still virtually nonexistent in the DACH region, would be premature; the data work from the table, on the other hand, pays off immediately because it serves the channels that matter today: Google Shopping, price comparison sites, and marketplaces—right in the middle of preparations for the holiday shopping season. Our sales platform comparison provides an overview. The practical recommendation: monitor the situation, build a database, and don’t rush into investing in individual protocols.
What the XONIC Shop System Offers for This
The data foundation for agents is standard in the XONIC Shop System: product pages contain structured product data as JSON-LD, the Google Merchant Feed is generated directly from the shop, and GTIN/EAN and MPN are standard product fields. Through the interfaces, 17 marketplaces and 14 inventory management systems can be integrated—inventory levels and prices remain consistent across all channels. The AI crawler control (since version 4.8.2) regulates which bots are allowed to crawl the store, and SEO URLs—including redirect history—keep product references stable. See the feature overview for a complete list of functions.
Frequently Asked Questions About Agentic Commerce
Sources
Stephan Dunger
Stephan Dunger is one of the brains behind the XONIC store system. He has been developing the platform together with the team since 2012 - from the database to the interfaces to the checkout.
A passionate programmer, technical mind and consultant at the same time: with his in-depth knowledge of store systems and e-commerce, Stephan combines the depth of a developer with an eye for the big picture. Together with the XONIC team, he shapes the technical direction, consistently thinks about functions from the retailer's perspective and advises on customized solutions.
The result is software with a face: customers don't get an anonymous provider, but a direct line to the people who develop XONIC. Pragmatic, fast and at eye level.
