Agentic Commerce Checklist: Get a Shopify Store Ready

What is agentic commerce, and what changes for a store?

In agentic commerce, a shopper asks an AI assistant for a product, and the assistant reads catalogs, feeds and store policies to answer. A person on your homepage sees your brand. An assistant sees fields, and a wrong field is a wrong answer.

OpenAI's commerce documentation describes the Agentic Commerce Protocol as an open standard that connects merchants to ChatGPT users and lets ChatGPT ingest structured catalog data. For you, the work is operational: every system that feeds those fields has to tell the truth. That is a different topic from agentic operations, where AI agents do your back-office work. Here the agents are the customer's, and your systems are what they read.

Which protocols matter to a Shopify merchant?

Three things matter today: the Agentic Commerce Protocol, the Universal Commerce Protocol, and Shopify's own agentic storefronts. Each is described only as far as its owner's page states it, and each changes often, so read the linked page before you act.

• Agentic Commerce Protocol (ACP): Stripe's documentation calls it an open standard created by Stripe, OpenAI and Meta. It defines agentic checkout, cart and feed, payment token delegation, authentication delegation, and order webhooks. OpenAI's pages say product feed onboarding in ChatGPT is currently open to approved partners, and recommend a full catalog snapshot at least once a day.
• Universal Commerce Protocol (UCP): the UCP site lists Google, Shopify, Etsy, Wayfair, Target and Walmart among its co-developers. Google's merchant page says UCP enables agentic actions on AI Mode in Google Search and Gemini, starting with direct buying, and that you stay the merchant of record.
• Shopify agentic storefronts: Shopify's help center says eligible products are discoverable by AI channels through Shopify Catalog by default, with Google AI Mode and Gemini fed through the Google & YouTube sales channel. Shopify's March 24, 2026 announcement names ChatGPT, Microsoft Copilot, AI Mode in Google Search and the Gemini app as channels.

How do you check your Shopify store's agent settings?

You check them in the Agentic section of your Shopify admin, which shows whether your catalog is shared, how many products are in it and why a store might be ineligible. Do this first, because Shopify's default setting, Allow Shopify to manage for me, shares your products and turns on direct checkout where supported, and it enrolls you in future agentic storefronts automatically.

Work through these steps and note what you find:

1. Open Sales channels, then Agentic. Note whether Shopify Catalog access is on, the product count, and any ineligibility message.
2. Decide on purpose whether to keep Allow Shopify to manage for me on. Turning it off lets you control catalog access and direct checkout channel by channel.
3. If your titles or descriptions live in custom fields, open the channel mapping page and confirm the right source is mapped. If you use compare-at prices, choose whether Catalog treats the price or the compare-at price as the list price.
4. Open yourstore.com/agents.md in a browser. Shopify serves it automatically, with /llms.txt and /llms-full.txt, and it carries your store name, sitemap link and store policies.
5. Check any B2B or wholesale setup. Shopify says agentic storefronts support only direct-to-consumer sales, and a hidden-price setup built with apps or theme code may not be detected as B2B-only.

Which product data should you fix first?

Fix the fields that Shopify lists as inputs to AI matching: title, description, images, type, vendor, collections, tags, barcode and variant option names. Shopify's guidance for AI platforms asks for clear, accurate and detailed information in those fields, and OpenAI's best practices ask for concise, factual product copy.

Turn that into a data-quality query instead of a one-off clean-up. Export your products and count the blanks in each field. Sort the blanks by vendor and by product type, so you can fix them in batches instead of one product at a time. Check three things in particular:

• Variants: OpenAI's guide asks for a stable parent product ID and a unique ID per purchasable option, with variant-specific titles, prices and availability where they differ. A variant named "Default" or a size buried in the title breaks that.
• Barcodes: Shopify lists the barcode (ISBN, UPC, GTIN and similar) among the fields AI platforms consider, so a blank barcode column is a gap you can count.
• Prices: a stale compare-at price turns into a false "was" price anywhere a channel shows strikethrough pricing. Review compare-at values whenever a promotion ends.

How do you keep stock and delivery promises accurate?

Keep them accurate by fixing the systems upstream of the storefront, because Shopify says Catalog keeps syndicating your latest inventory and prices. If the stock count in Shopify is wrong, every AI channel inherits the error, and the order lands in your warehouse as a problem. Our guide to phantom stock and open purchase orders shows how to compare Available against what is physically sellable.

For delivery promises, publish only a window you actually meet. Google's merchant listing markup lets you state shipping cost, destination, handling time and transit time as structured fields. Fill them from your own records:

• Pull the last 90 days of fulfilled orders and measure days from order to delivery for each destination.
• Pick the window that covers most orders for that destination, with a day of margin for busy weeks.
• Compare it with the promise on the product page, in checkout and in your shipping policy. Three different numbers is a finding.
• Re-measure monthly, and after any carrier or warehouse change.

How do you turn returns terms and reviews into data?

Write your return terms as fields, not only as a page of paragraphs. Google's MerchantReturnPolicy markup takes a return window in days, a return method, return fees and the countries the policy covers, and it nests under your organization markup. Shopify also advises keeping store policies complete and up to date so an AI agent that is asked about returns can cite what you stored.

Then make the terms match what your team does. If support routinely accepts returns after the stated window, the window is the wrong number. Write the real rules into one SOP and publish from it.

Reviews follow the same logic. Google says customer reviews and ratings can enhance how a product appears in Search, and that enhancements show at each experience's discretion. Check that your review app outputs structured review data on product pages, and trigger review requests from the delivered status rather than the order date, so customers rate the product they have in hand.

What should you measure once agents send orders?

Measure three things: where orders come from, how many of them hit a problem, and how fast problems close. Shopify says orders from ChatGPT arrive in the admin with referral attribution, so you can count them separately from your other channels.

Track the share of those orders that end in a support ticket, a cancellation or a return, and compare it with your other channels. A higher rate points to a field an agent read that was wrong: a promise, a variant, a price. Fix the source field, then re-check the rate. A diagnosis of your operations uses the same logic across stock, support and fulfillment.

The checklist above needs no new tool, only settings and exports you already have. If you later want agents working on your side of the business, agent workflows are the written procedures they follow.

Sources

Every protocol and platform fact above was read on the owner's own page on November 11, 2026. Pages and settings change, so check the page before you rely on a step.

• OpenAI, Commerce overview, Get started and Best practices.
• Stripe, Agentic Commerce Protocol and Universal Commerce Protocol.
• UCP, ucp.dev and Google, Getting started with the Universal Commerce Protocol.
• Shopify newsroom, Millions of merchants can sell in AI chats, March 24, 2026.
• Shopify Help Center: Managing agentic storefronts, Shopify Catalog and product discovery, Optimizing your products for AI platforms and Shopify Catalog data mapping.
• Google Search Central: Product structured data, Merchant listing structured data and Merchant return policy structured data.

Frequently asked questions

Do I need to build an ACP or UCP integration myself on Shopify?

Usually not. Shopify's March 24, 2026 announcement says products become discoverable in ChatGPT by default through agentic storefronts, with no separate integrations or apps, provided the store is eligible for Shopify Catalog. A store on another platform has to follow each protocol's own documentation. Shopify also offers an Agentic plan for brands that do not sell on Shopify and want their products in Shopify Catalog.

Can I opt out of AI channels?

Yes, in part. Shopify's help center says you can turn off Shopify Catalog access, and that it can take up to 7 days before your data stops being shared. Your products may still be found through web crawling and indexing, the same way a search engine finds them. Setting a product's status to Unlisted hides it from AI channels completely.

Does traditional SEO still matter for agentic commerce?

Yes. Shopify's help center tells merchants to keep product information complete and well structured in Shopify Catalog and to follow SEO best practices, and it notes that AI crawlers can also reach a store through the open web. Accurate titles, descriptions, images and policies serve both.