# Shopify Lets AI Agents Read, Edit and Submit Checkout on Its Own > Source: > Published: 2026-09-29 02:19:31+00:00 *Shopify just gave AI agents the same access to checkout that a logged-in shopper has: read the order, change the address, hit submit.* On September 28, Shopify's developer changelog quietly posted an entry that matters more than its title lets on: "WebMCP support for checkout." The update hands AI agents three new tools, get_checkout, update_checkout and complete_checkout, that let them inspect a shopper's cart, change the delivery address or shipping option, and place the order once the buyer signs off. TechCrunch reported the rollout covers all eligible Shopify merchants, including checkouts that run through Shop Pay. That's a meaningfully different thing from what Shopify shipped in August. The earlier WebMCP rollout let an agent browse a catalog and build a cart inside a storefront. It could window-shop. It could not check out. This update closes that last gap, and it does it without screen-scraping or screenshots, the two hacks every prior agent-shopping demo has leaned on. WebMCP isn't a Shopify invention. It's a proposed browser standard being built jointly by Google and Microsoft inside the W3C Web Machine Learning Community Group, and it's currently running as a public origin trial in Chrome, with early support in Microsoft Edge. The idea is simple: instead of an AI agent guessing at a page's layout by parsing pixels or DOM elements, a storefront registers its own functions directly with the browser, the same way a page registers a click handler today. The agent calls the function. The browser executes it in the shopper's own session, with the shopper's own cookies and payment methods, not some separate server-side proxy account. That distinction is the whole story. Shopify already runs a hosted Model Context Protocol server that lets agents talk to a store server-to-server. WebMCP is different because it lives inside the buyer's browser tab, tied to a real, authenticated shopping session. Gil Greenberg, a Shopify staff product manager working on agentic commerce, said on X that the checkout tools are rolling out to all eligible merchants now, no opt-in flag, no beta waitlist mentioned. [Shopify lets AI agents finish checkout on its two million stores](https://startupfortune.com/shopify-lets-ai-agents-finish-checkout-on-its-two-million-stores/) Shopify expanded its WebMCP protocol on September 28 so AI agents like Meta's Muse and Instinct can complete checkout, not just browse and add to cart, across its more than two million merchant stores. The move outruns any settled answer on who is liable when an autonomous agent, not a human, submits the order. - [AI agents completing checkout on Shopify stores](https://startupfortune.com/shopify-lets-ai-agents-finish-checkout-on-its-two-million-stores/) - [how Shopify WebMCP agents submit orders automatically](https://startupfortune.com/shopify-lets-ai-agents-finish-checkout-on-its-two-million-stores/) Shopify is also leaning on the Universal Commerce Protocol, an open standard for how agents discover products, build carts and check out across any platform or payment processor. UCP's backer list reads like a truce among rivals: Amazon, American Express, Etsy, Mastercard, Meta, Microsoft, Salesforce, Stripe, Target, Visa and Walmart are all listed as supporters, according to Shopify. ## Amazon is doing the opposite Here's what makes Shopify's timing sharp. Just days earlier, Amazon blocked Meta's Muse assistant from shopping on Amazon.com on customers' behalf. GeekWire reported Amazon tried first to get Meta to voluntarily exclude Amazon from Muse's shopping flow, and when that failed, cut it off outright. Amazon's stated reasons, per its own account: Meta never disclosed that Muse would access the store, the agent doesn't identify itself while browsing, and it appeared to capture and store customer credentials in a way Amazon says creates privacy and security risk. Amazon had already written a formal Agent Policy into its Business Solutions Agreement back in March, and Adidas has taken a similar blocking stance on its own site. So you've got two philosophies running side by side in the same month. Amazon treats an unidentified agent acting inside its checkout as a security threat until proven otherwise. Shopify is opening the door wide and betting that identifying the agent, keeping it inside the buyer's own authenticated browser session, and requiring explicit buyer authorization before complete_checkout fires, is enough guardrail to make the risk worth the traffic. For a merchant running a Shopify store, that bet lands directly on them, not on Shopify. If an agent like Meta's Muse or Instinct completes a purchase on your storefront with WebMCP now live, that's a real order, real revenue, no different from one placed by a person typing into a browser. But it also means your storefront's product data, pricing logic and inventory state now have to be accurate enough to survive an agent reading them programmatically instead of a human squinting at a product page and forgiving a typo or a slow-loading image. Sloppy variant data or broken inventory sync, previously a minor annoyance, becomes something an autonomous purchasing agent can act on badly at scale. Frankly, the merchants who benefit most here are the ones who already treat their product feed as a structured asset rather than an afterthought. The ones who don't are about to find out the hard way what an agent does with ambiguous data it was never designed to second-guess. **Also read:** [Cognition cuts Devin's price and raises its benchmark score in the same week](https://startupfortune.com/cognition-cuts-devins-price-and-raises-its-benchmark-score-in-the-same-week/) • [TSMC Is Quietly Setting the Price Every AI Chip Company Has to Pay](https://startupfortune.com/tsmc-is-quietly-setting-the-price-every-ai-chip-company-has-to-pay/) • [Shopify lets AI agents finish checkout on its two million stores](https://startupfortune.com/shopify-lets-ai-agents-finish-checkout-on-its-two-million-stores/) [Shopify CEO Tobias Lütke Warns His Own AI Mandate Is Creating Slop Grenades](https://startupfortune.com/shopify-ceo-tobias-ltke-warns-his-own-ai-mandate-is-creating-slop-grenades/) Shopify CEO Tobias Lütke used a Knowledge Project podcast interview to coin the term "slop grenades," describing unreviewed AI-generated drafts that dump verification work on coworkers. The comments land a year after Lütke made AI use a baseline expectation at Shopify, where roughly half of production pull requests now start as Slack conversations... - [shopify ceo warns about ai generated slop grenades](https://startupfortune.com/shopify-ceo-tobias-ltke-warns-his-own-ai-mandate-is-creating-slop-grenades/) - [mandatory ai use creating unreviewed draft problems](https://startupfortune.com/shopify-ceo-tobias-ltke-warns-his-own-ai-mandate-is-creating-slop-grenades/) *This article is posted in [Technology News](https://startupfortune.com/category/technology/), check it out for more related stories.* ## Join the discussion [Open in the community →](https://startupfortune.com/community/) Almost there. Sign in and your reply posts straight away.