Shopify just shipped something most merchants have not noticed yet. Buried inside the broader agentic commerce infrastructure rollout is a free tool at commerce-readiness.shopify.io that takes any store URL and scores how readable that store is to AI shopping agents like ChatGPT, Microsoft Copilot, and Google AI Mode.

You put in your URL. It runs an analysis. It tells you exactly what AI agents can and cannot see when they evaluate your store for a purchase recommendation.

Gymshark, one of the most technically sophisticated DTC brands on Shopify, scored 84%. They passed 13 checks and failed a few. The gaps were not obvious ones. Their reviews exist on product pages. Their shipping policy is in the footer. Both look complete to any human visitor. Neither counts to an AI agent unless it is structured in machine-readable format.

That gap between what your store shows humans and what it communicates to AI agents is the problem this tool was built to expose. And for most Shopify merchants, the score is going to be lower than expected.

This article covers what the tool actually checks, why stores are failing the checks they are failing, what the underlying infrastructure shift means for how merchants compete in 2026, and what to prioritize first. It is not a surface-level overview. Shopify's field CTO Sandy Jeong released a detailed technical guide alongside this tool. What follows draws from that guide, from the structured data research published by Ecommerce Fastlane, from Shopify's own agentic commerce documentation, and from the technical breakdowns published by Presta and Ask Phill.

What Shopify's Agentic Commerce Readiness Tool Actually Is

Shopify activated agentic storefronts for all eligible merchants in late March 2026 as part of the Winter '26 Edition. The short version: your products are now automatically discoverable inside ChatGPT, Microsoft Copilot, and Google AI Mode if your store is in the Shopify Catalog and you sell to US shoppers. Orders flow back into your Shopify admin with full channel attribution. You did not opt in. This is on by default.

The readiness tool is Shopify's acknowledgment that being discoverable is not the same as being competitive. Your products can be in the Shopify Catalog and still be invisible to the AI agents that are supposed to surface them, because the agents cannot parse your data well enough to recommend you confidently over a competitor whose data is cleaner and more complete.

The tool scores your store across three categories. Agent Discovery measures whether AI crawlers can find, index, and understand your store. Product Intelligence measures whether your product data is structured in a way agents can actually use to make recommendations. Transaction Readiness measures whether the purchase flow, policies, and trust signals are machine-readable, not just human-readable.

Gymshark's breakdown from the tool screenshot that circulated: Agent Discovery at 93%, Product Intelligence at 70%, Transaction Readiness at 90%. Their weakest area was product intelligence, which is the category most Shopify merchants will score lowest on. The reasons why matter.

The Structured Data Gap: Why Your Store Looks Complete and Is Not

When a customer lands on your product page, they see your photography, your brand story, your reviews, your shipping information. They process all of it the way a human does, reading copy, looking at images, scanning the footer for policies.

An AI shopping agent sees none of that in the same way. It parses your structured data. It reads your JSON-LD markup. It processes your API responses. Everything else, the visual design, the lifestyle photography, the scroll animations, is either invisible to the agent or irrelevant to its decision-making process.

The problem is that Shopify's default themes generate between 5 and 8 structured product attributes. The title. The price. The description. The image URL. Availability. Maybe a category tag. That is what your store communicates to an AI agent by default.

AI agents making purchase recommendations need 30 or more structured attributes to operate confidently. Ecommerce Fastlane's guide documents the gap precisely: AI agents need material composition, size specifications with actual measurements, care instructions, compatibility details, occasion tags, sustainability certifications, GTIN identifiers, aggregate ratings in structured format, return policy in schema markup, and shipping estimates in machine-readable fields.

If an agent cannot determine from your structured data whether your jacket is waterproof, whether your supplement is vegan, or whether you accept returns within 30 days, it will recommend a competitor whose data answers those questions. The competitor does not need a better product. They just need cleaner data.

Gymshark lost points on Product Intelligence for exactly this reason. Their reviews are on the product page as rendered text. Humans can read them. AI agents need reviews in JSON-LD format with AggregateRating schema to parse them as a trust signal. Their shipping policy is in the footer in plain text. It is findable in two seconds by any shopper. An AI agent cannot reliably extract it from unstructured text. It needs MerchantReturnPolicy and ShippingDetails schema injected into the product JSON-LD to count it.

This is not a fringe technical issue. This is the core infrastructure problem the readiness tool was built to surface. And it affects every Shopify store that has not specifically addressed structured data beyond Shopify's defaults.

The Five Things the Tool Checks and Why Each One Matters

LLMs.txt

This is a plain text file placed at the root of your domain, at yourdomain.com/llms.txt. The emerging standard, documented by Alhena AI, tells AI crawlers what your site offers, where to find key information, and what sections are most relevant for commerce interactions. It is the equivalent of robots.txt but for language model crawlers instead of search engine bots.

Most Shopify stores do not have one. The stores that do have a meaningful advantage: AI crawlers can navigate their site architecture efficiently rather than inferring structure from HTML. For merchants on Shopify's default Liquid themes, adding an llms.txt file is one of the fastest wins available because it requires no app installation and no developer work beyond creating and uploading a structured text file.

Breadcrumb Schema

Breadcrumb schema is machine-readable navigation structure that tells AI agents where any given page sits within your catalog hierarchy. A product page for a men's waterproof running jacket communicates to a human through the visual navigation. It communicates to an AI agent through BreadcrumbList schema: Home > Men's > Running > Jackets > Waterproof.

Without breadcrumb schema, agents have to infer category relationships from URL structure and page content, which introduces error. With it, agents can accurately understand what you sell, how your catalog is organized, and whether a specific product matches a shopper's query. The Weaverse technical breakdown from Shoptalk 2026 specifically calls breadcrumb schema one of the three infrastructure pieces merchants should activate immediately.

FAQ Schema

FAQ schema is structured markup that tells AI agents what questions your brand answers and what those answers are. The readiness tool checks whether your about page and product pages communicate brand context, shipping information, return policies, and product specifications in machine-readable FAQ format.

The practical issue: most merchants write this information as flowing copy. Your about page explains your founding story in marketing language. Your product pages describe your materials in brand voice. Both are well-written for humans. Neither is structured in a way an AI agent can extract and use reliably when answering a shopper's question.

Ask Phill's analysis makes the principle explicit: the phrase "100% GOTS certified organic cotton, 200 GSM" consistently outperforms "luxuriously soft premium cotton" in every agent ranking algorithm, because the first is a factual specification the agent can parse and match to a query, and the second is marketing copy the agent has to interpret.

JSON-LD Product Markup

This is the most important check the tool runs and the area where most stores score lowest. JSON-LD is the structured data format that tells AI agents everything they need to know about a product in a machine-readable layer that sits alongside the human-readable product page.

Shopify's default themes inject basic Product JSON-LD: name, price, image, availability, maybe a currency field. The gaps are significant. GTIN and EAN identifiers, which AI agents use to cluster products and compare across merchants. AggregateRating schema for reviews, without which agents cannot use your social proof as a ranking signal. MerchantReturnPolicy schema, without which agents cannot confirm you accept returns and under what conditions. ShippingDetails schema with actual delivery estimates. ProductGroup schema for products with variants.

Shopify's own agentic commerce documentation is clear about why this matters: AI agents do not browse your product pages the way humans do. They rely on structured data. Product information like title, price, material, and dimensions must be organized in standard, machine-readable fields rather than embedded in marketing copy or page layouts. If the data is not there in the right format, the agent cannot use it.

The Redlio Designs technical breakdown frames the stakes correctly: in 2024, brands blocked AI bots to protect their content from training data. In 2026, those same bots are the primary shoppers. Blocking GPTBot or ClaudeBot is the equivalent of de-indexing your site from Google. The stores that treated AI crawlers as threats in 2024 are now invisible to AI agents in 2026.

Review and Rating Schema

The fifth check the tool runs is specifically on review and rating data. This is the check that caught Gymshark. Reviews in plain text on a product page are human-readable but agent-invisible. Reviews in AggregateRating JSON-LD format are machine-readable and function as a trust signal AI agents actively use when ranking products for recommendations.

For merchants using review apps like Loox, Judge.me, or Okendo, the question is whether those apps are outputting structured AggregateRating schema or just rendering star ratings as HTML. Most do output schema markup, but the configuration matters. An average rating of 4.8 from 2,400 reviews is a powerful trust signal. If it is not in structured format, the AI agent making a recommendation cannot use it.

The Protocol Layer Behind the Tool

Understanding why Shopify built this tool requires understanding the protocol infrastructure that now underlies agentic commerce. There are three protocols merchants need to know, even at a surface level.

Shopify's Storefront MCP configures your Shopify Storefront API as an MCP endpoint, which lets AI agents query real-time product data, manage carts, and guide checkout through direct API interaction rather than through crawling. For merchants on Hydrogen, Shopify's headless React framework, this is native. For merchants on Liquid themes, MCP support exists but with structural limitations, Liquid themes are built for browser rendering, not optimized for machine-readable API responses.

OpenAI's Agent Commerce Protocol was originally built with Stripe and aimed at native ChatGPT checkout. OpenAI scaled back from direct checkout in March 2026, shifting to a referral model where ChatGPT links to your store rather than completing the purchase inside the chat interface. Your product data in ChatGPT's index still determines whether you get recommended. The checkout happens on your store. The discovery happens inside the AI.

Anthropic's Model Context Protocol, now under the Linux Foundation, provides the general framework for AI-tool interaction that is increasingly relevant as Claude and other assistants add shopping capabilities. For Shopify merchants, the practical takeaway from all three protocols is the same: your product data feeds into multiple AI systems simultaneously. Fixing your structured data once serves all of them.

What Merchants on Liquid Themes Can Do Without a Developer

Most Shopify merchants are on Liquid themes, not Hydrogen. The structured data gap is real for Liquid stores, but it is not insurmountable. There are specific actions available without touching code.

The first is adding an llms.txt file to your store root. This requires creating a structured text file and hosting it at yourdomain.com/llms.txt. Shopify does not have a native UI for this, but it can be done through a custom app file or by working with your theme developer to add it as a static asset. The file tells AI crawlers your store's name, what you sell, your key page URLs, and your catalog structure.

The second is auditing your Google Merchant Center feed. Ask Phill's guide documents that Google's Shopping Graph feeds multiple AI agents, not just Google's own. Updated guidelines from NRF 2026 recommend product titles of 30 or more characters, descriptions of 500 or more characters, GTIN always populated, and a minimum of three additional product images per listing. Your Merchant Center feed quality directly affects your visibility in Google AI Mode.

The third is using a JSON-LD enhancement app. Shopify Growth Services recommends implementing JSON-LD Product schema on every product page with full attribute mapping, including Offer schema with real-time price, availability, and currency; ProductGroup schema for products with variants; Brand, AggregateRating, and Review schema where data exists. Apps like JSON-LD for SEO handle much of this without developer work.

The fourth is auditing your product catalog attribute depth. Export your product catalog and count the structured attributes per product. If you are averaging fewer than 10 attributes per product, you are invisible to most AI agent comparisons. The target is 30 or more: material, dimensions, care instructions, compatibility, occasion, certifications, GTIN, variant-level detail. This is a content and data task, not a technical one, but it is the highest-leverage action available for improving Product Intelligence scores.

The fifth is reviewing your policy pages for schema eligibility. Your return policy, shipping policy, and FAQ content should be in structured format. Review apps and policy apps vary significantly in whether they output machine-readable schema. Check what your current apps are outputting before assuming the structured data exists.

The Competitive Window Is Short

Most Shopify merchants do not know this tool exists. The LinkedIn post that surfaced it this week came from a founder running the tool on major brands for the first time. The structured data work it points to has been discussed in technical circles for months, but merchant adoption is low.

The window to get ahead of this is real and it is narrow. The merchants who address their structured data gaps now will accumulate a compounding advantage: AI agents learn which merchants provide reliable, complete product information over time. Early compliance builds the kind of indexed trust that late adopters cannot buy their way into quickly. Ask Phill's analysis is explicit on this: the structured data, feed quality, and crawler configuration work done today compounds over time as AI agents learn which merchants provide reliable, complete data.

The fact that Shopify shipped a compliance scoring tool alongside the agentic storefront rollout is not a coincidence. It is Shopify telling merchants directly that readiness is not automatic. Being in the Shopify Catalog is the floor. Structured data quality is what determines the ceiling.

Run your store through commerce-readiness.shopify.io before your competitors do. The score will tell you exactly where the gaps are. The fixes are documented. The timeline to act is now.

Where Order Operations Fit Into This

There is one dimension of agentic commerce readiness that the tool does not score and that most of the technical coverage has not addressed: what happens to the orders AI agents place.

AI agents do not place orders the way human shoppers do. They act on data. They follow structured purchase flows. They do not double-check their own inputs. When an AI agent completes a purchase on behalf of a shopper, the order arrives in your Shopify admin exactly like any other order. But the address confirmation, the fraud signal check, the order detail verification that a careful human shopper might catch themselves, none of that happens in the same way.

The 2.1% bad address rate documented by Shippo for e-commerce parcels was measured against human-placed orders. As AI agent-placed orders become a material percentage of Shopify transaction volume, the operational question is not just whether agents can find your store. It is whether your order operations are ready to handle what agents send you.

Tacey sits at this layer. The moment any order is placed on your Shopify store, whether by a human shopper or an AI agent acting on their behalf, Tacey checks the address, reads the order signals, and makes a decision before your warehouse sees it. PASS means the order moves. AUTO-RESOLVE means Tacey fixes a correctable issue automatically. FLAG means the order needs human review before it ships. The Operations Toolkit layers on top of that with fulfillment management, discount intelligence, and customer analytics across your full order history. At $39 a month on the Scout plan covering 500 orders, the carrier correction fees a single prevented bad shipment avoids typically cover the monthly cost.

Agentic commerce readiness is a two-part problem. The first part is discoverability: whether AI agents can find you, read your data, and recommend you. The tool at commerce-readiness.shopify.io addresses that part. The second part is operational: whether your order handling is ready for the volume and the order profile that agentic commerce generates. That part is still largely unaddressed in the mainstream conversation.

Both matter. Fix your structured data. Run your score. And make sure the orders that come in are handled before they become the kind of delivery failure that ends a customer relationship an AI agent worked to create.