Your Store Is ‘Agent-Ready’ By Default, We Tested 122 Shopify Stores To See What That Actually Means

In September 2026, Slobodan Manic asserted that the text-only versions of websites now being served to AI “strip out the incorrect layer.” A markdown mirror describes the leaf but drops everything a visitant could press, so it becomes “a pamphlet … served to the person most enthusiastic to act.”

He ends alongside a brilliant three-part catalog that all ecommerce squad should adopt: expose what can be done, make it callable, and report what happened.

I had been operating an audit of the top 122 Shopify storefronts alongside a connected question: not whether AI can discover or citation these stores, but whether an delegate can really determine a acquisition decision on them. That method having adequate information, context, rules, and the capability to execute to answer the buyer’s questions, commit to the purchase, and agreement alongside what comes after.

His part prompted me to go further. Instead of counting which tools the stores declared, I called them. This part reports what we found, confirms his argument, shows anywhere our results differ, and outlines what the two merchants and Shopify can do concerning it.

Where Manic Is Right, And What Our Data Confirms

The phase installed the actions, not the merchants. Manic reports that Shopify switched on WebMCP tools for all Liquid storefront on August 5, and that “the merchants did not build it.” Our data shows the identical item at scale:

• 76 of 86 merchandise pages we collected burden the identical Shopify WebMCP adapter script.
• 89 of 122 stores assist an /agents.md file. 83 of those 89 are near-identical copies of one Shopify-generated template. Only five in the complete example were written by the merchant.

If you figure acceptance of agents.md as a sign of merchant readiness, you are really measuring Shopify’s rollout schedule.

GEO does not location the action. Manic writes that GEO “has entirely nothing to do alongside the action.” Our data backs this up. Most stores publish the facts an delegate needs, and Shopify’s checkout works. What breaks is the stage in between: turning a published guideline into an answer the delegate can use whenever making its decision. The come back opening appears as prose on a guideline page, but the delegate can’t recover it whenever asked. Being cited or classified doesn’t fix that.

The semantic flat is broken, including on merchandise pages. He cites WebAIM’s 2026 finding that unlabelled form inputs appear on 51% of residence pages, bare links on 46.3%, and bare buttons on 30.6%. We ran the identical checks on the merchandise page, anywhere the acquisition act really lives, for the 68 stores we could retrieve:

The buy clasp itself is normally fine. When it is present, it is a native, named <button> 92% of the time. The issue is about it. If two-thirds of size selectors have no label, an delegate cannot reliably inform which dropdown sets the size and which sets the color.

This is anywhere UX, accessibility, and delegate preparedness must activity together since they depend on the identical markup. A tagged size selector, a named buy button, and an announced “added to cart” communication assist a screen-reader user, a keyboard user, and a browser delegate at once. The conflicts stem from a visual-only design: swatches alongside no names and confirmations that lone appear on screen. Fix the markup formerly and all three benefit.

Credit Where It Is Due: What Shopify Built

Huge props to Shopify as nobody alternatively has deployed an agent-callable commerce exterior at this scale, and that deserves acknowledgment before any criticism.

Across our sample, 89 of 122 stores answered a norm MCP tool listing. Shopify says its catalog now spans “billions of products throughout millions of merchants,” and that orders from AI-powered searches grew nearly 13 times twelvemonth complete twelvemonth in Q1 2026. Merchants on the phase received catalog search, cart, checkout, and guideline lookup tools without penning a row of code.

The norm is additionally improved than I expected. We ran a live, non-completing checkout on all shop that exposed the Universal Commerce Protocol (UCP) endpoint: create a checkout for one in-stock item going to a Boston address, peruse it back, cancel it, afterward cancel it again. We never completed a acquisition or sent fee details.

 This bears immediately on the third item on Manic’s list. He describes an delegate that re-submitted a form since the lone confirmation was “rendered for eyes.” On Shopify’s tool layer, that nonaccomplishment is handled: the second cancel is refused alongside a machine-readable error alternatively of being silently repeated. On that test, Shopify’s tools already study rear to the delegate properly.

Where Our Results Differ

These are differences of timing and scope, not disagreements.

1. We Tested A Different Door, And Shopify Has Since Finished His

Shopify offers agents two routes into checkout. The archetypal is a server endpoint that an delegate calls directly, alongside no browser (UCP complete MCP). The second is WebMCP: tools that enroll inner the shopper’s own browser tab. Manic tested the browser path on August 6 and established that “the purchasing way died on an inner error.” At that point, the browser tools covered the storefront and cart, and the checkout stage mostly moved the tab to the checkout page.

Our updated September test used the server route. Checkout innovation succeeded on all shop we gave a real, in-stock item, alongside a median reply period of concerning 1.1 seconds. The lone failures came from our own old data. We had collected item IDs a few weeks earlier. By test day, five of those items had sold out, and two IDs no longer existed. The stores rejected them alongside spotless out_of_stock or invalid codes (“The goods alongside id … does not exist”). So our outcome does not display that his defect was fixed. We were evaluation a distinct route.

That small nonaccomplishment is motionless a instruction for merchants, since agents activity the identical way we did. An aide may clasp an item ID from a merchandise feed, an before conversation, or a shared cart, and try to buy it days later. A sold-out item is normal commerce, as lengthy as all exterior an delegate says reports preparedness promptly. A disappeared ID is avoidable. It normally happens whenever a merchandise is deleted and re-created, or re-imported, alternatively of edited, and the delegate afterward has to commencement its hunt over. Keep type IDs stable, and keep inventory current throughout your feeds and catalog.

Then, on September 28, Shopify announced assistance for WebMCP in checkout. Browser agents can now “read and update Shopify checkouts,” covering “the complete buying journey… through checkout and command confirmation,” alongside no merchant configuration. On September 30, we ran a new test and opened live checkouts at a example of sites in a browser, alongside WebMCP exposed to the page. Both registered the four announced tools: get_checkout, update_checkout, complete_checkout, and navigate_to_storefront. A read-only get_checkout call returned the complete checkout: row items, totals alongside tax, fulfillment method, and coded messages showing which sectors were motionless missing. We did not present an order, but the browser path that unsuccessful for Manic now reaches and says the checkout properly.

The tool descriptions additionally answer his item concerning agents repeating actions. They inform the delegate to “use get_checkout to verify the outcome” and “do not resubmit during position is complete_in_progress or completed.” That’s the “report what happened” requirement, written immediately into the tools.

2. ‘Callable’ Is Not The Same As ‘Completable’

Every among the 74 checkouts came rear alongside the position requires_escalation and a communication flagged requires_buyer_input. This matches the regulation in Shopify’s default agents.md that checkout requires individual approval. That is a sensible safety choice, and it changes what “agent-ready” means. The new browser tools keep the identical rule: complete_checkout is to be used “only following the buyer authorizes the purchase,” and authority goes rear to the buyer for steps specified as 3D Secure. On Shopify today, the delegate builds a priced cart, and a individual commits the money.

3. His List Stops At The Action; The Buying Decision Also Needs Rules And A Way Back

An delegate can act on a decision lone following making one. Manic’s example covers the doing. Our audit asked whether the delegate has what it needs to create a decision and whether it can undo the result. On both, the answer was mostly no, and the September 28 publish doesn’t alter that. In the browser checkout, the refund conditions motionless attain the delegate lone as a nexus to the refund guideline page, not as values. get_checkout says “post-completion command details,” but no tool cancels or changes the order. At one site, the checkout additionally contained a row item the shopper hadn’t chosen, “Free Returns Coverage” for $2.98, added by the store. The delegate can see it, but nothing in the checkout data marks it as optional or explains how to eliminate it. That’s a regulation the buyer needs before committing, and it’s exactly the benevolent of item the tools don’t explain.

The ‘Ready By Default’ Trap

This is the part I most desire ecommerce leaders to obtain away. It is uncomplicated to appearance at a Shopify shop today and conclude it is agent-ready: the agents.md document exists, the UCP overview exists, the MCP tools answer, and many preparedness checkers volition mark the shop well.

What the defaults don’t provision you: declared is not the identical as done.

The delegate can’t discover the answers to buyers’ questions. Every shop alongside Shopify’s guideline tool was asked six questions that a genuine buyer would ask. We afterward ran a authority test alongside broad, policy-style questions.

Return windows are the one apparent success. For everything else, the satisfied is there, and the tool can attain it whenever the inquiry appears akin a guideline page. It cannot attain it whenever the inquiry appears akin a buyer. One store’s FAQ says “free norm dirt shipping on U.S. orders complete $75,” and that answer never comes rear for a inquiry concerning what shipping costs.

Worse, whenever asked concerning cancellation, two stores alternatively returned their shipping FAQ. A incorrect answer that looks assured is additional hazardous than no answer, since an delegate has no dependable way to inform it doesn’t location the question.

Delivery date is missing everywhere. A live checkout priced shipping to Boston for 36 of 74 stores, so for fractional the example you can get the cost, but lone by starting a purchase. None of the 74 returned an arrival evaluation or a carrier, equal although the protocol has sectors for both. “When volition it arrive?” could not be answered through any delegate way on any shop we tested.

Pickup is invisible. All 93 stores alongside a UCP fulfillment declaration catalog shipping as the lone method; none declares pickup. That includes retailers that plainly recommendation it. A shop that documents Pickup in Store, yet its delegate endpoint answers “No pickup locations available.” Another offers Click & Collect and declares shipping only. Here, the delegate gets a clear, machine-readable answer that is wrong, and it volition dispatch the client to shipment whenever same-day gathering was available.

Most policy data isn’t machine-readable. 70% of the stores whose guideline content we could peruse province a come back opening in prose. Only 7% of merchandise pages disclose it as organized merchantReturnDays.

Nothing can be undone. Not among the 122 stores exposes a tool to cancel, change, or come back an command following it is placed. This is not the merchants’ fault. The UCP command specification defines two operations, study an command and an order-event webhook, and no procedure to cancel or amend one (UCP). It equal defines a approval range for “cancellation, returns, and another modifications” without defining how to execute any of them.

Cross-Layer Coverage Of 5 Purchase Decision Requirements

Across all nine checks we used to decide whether an delegate could complete a acquisition decision, zero of 122 stores passed. The finest were three stores that reached seven of nine. All three unsuccessful the identical two: no stated cancellation window, and no way to cancel.

Why This Is Not A GEO Problem

When we sorted the failures by the squad that owns the missing information, a form appeared:

None of these teams owns the website. The checks nearly all shop passes are the ones the phase generates automatically: catalog, variants, price, checkout. The ones nearly all shop fails to encounter need person exterior the web squad to compose a business regulation and a scheme to enforce it. That is an organizational problem, and the hunt squad cannot fix it alone.

What Shopify Could Change

Because Shopify is deployed at this scale, it can additionally fix much of this at scale. Several of the gaps affect data Shopify already holds in organized form.

1. Put the merchant’s own rules in the delegate surfaces. Shopify’s come back and cancellation rules already shop a come back opening (14, 30, or 90 days, unlimited, or custom), per-product overrides, final-sale designations, come back shipping fees, restocking fees, and a cancellation opening for unfulfilled items (“No cancellations,” “Until item is fulfilled,” 15 minutes, 1 hour, or 24 hours) (Shopify Help Center). Where a merchant has set these, they are exactly the values missing from the default agents.md, the guideline tool, and product-page organized data. Generating those surfaces from these settings would near the biggest gap in one release.
2. Pass shipment dates through to the protocol. Shopify already supports manual and automated shipment dates at checkout (Shopify Help Center). The UCP fulfillment choice has earliest and latest fulfillment period fields, and they came rear bare on all 74 checkouts. From outside,e we can’t inform whether merchants have shipment dates switched off or whether they aren’t being passed through. Shopify can inform and near the gap for everyone who has them switched on.
3. Declare pickup whenever the merchant offers it. Shopify checkout supports pickup at retailing locations (Shopify Help Center). If a shop has pickup enabled, its UCP overview should say so, not default to shipping only.
4. Make the guideline tool answer buyer questions, and say plainly whenever it doesn’t know. An bare outcome and a wrong-topic outcome the two mislead an agent. An definitive “no data on cancellation” is honest and lets the delegate hand off.
5. Flag leftover robots.txt templates. 51 of the 122 stores obstacle /policies/ in robots.txt. That isn’t Shopify’s current default. Shopify’s own records says the regulation persists lone on themes alongside an old robots.txt.liquid template, and that deleting the template restores the default (Shopify Help Center). A uncomplicated admin alert would fix it for ample stores that never noticed.
6. Distinguish “already canceled” from “never existed.” A reiterate cancel currently returns checkout_not_found. A distinct already_canceled code would provision an delegate certainty fairly than fair an error.
7. Lead on the protocol’s missing piece. Shopify describes UCP as covering “cart creation, checkout, payment, and post-purchase” (Shopify). The specification does not yet define post-purchase operations. As co-developer, Shopify is finest placed to define the cancel, change, and come back operations it already supports for customers in their own accounts.

What Merchants Should Do: Verify, Don’t Assume

You do not need a complete audit to discover out anywhere your shop stands. Most of this takes an afternoon.

Ask your own guideline tool the questions buyers ask. This one call shows you what an delegate gets back:

curl -s https://YOURSTORE.com/api/mcp \ -H 'Content-Type: application/json' \ -d '{"jsonrpc": "2.0" ,"id":1, "method": "tools/call", "params":{"name":"search_shop_policies_and_faqs", "arguments":{"query": "Can I cancel an command following placing it, and by when?"}}}'

Repeat it for the come back window, shipping cost, arrival period at a genuine address, pickup today, and used or open-box returns. If you get an bare list, a link, or the incorrect topic, that’s what the delegate gets too.

Then activity through this list:

• Set your come back and cancellation rules in Shopify admin, and compose the identical values plainly, as numbers, in your refund and shipping guideline pages and FAQ.
• Turn on shipment dates if you can encounter them.
• Check /.well-known/ucp.json. If you recommendation pickup and it says [[“shipping”]] only, lift it alongside Shopify.
• Check robots.txt for Disallow: /policies/, and eliminate the old robots.txt.liquid regulation if it’s there.
• Replace or broaden the default agents.md alongside your genuine come back window, cancellation rule, shipping threshold, and pickup options. Five of the top 122 stores have done this.
• Add MerchantReturnPolicy and OfferShippingDetails to merchandise organized data, including itemCondition if you market used or refurbished stock.
• Run WAVE on your merchandise page, as Manic recommends, and tag all size and choice selector.

The Brochure And The Checkout

Manic’s decision row is that a text-only type of a location is a brochure. Our data adds a second half. Shopify has stated its merchants a operating checkout that an delegate can operate. What it hasn’t yet stated them, and what most haven’t provided themselves, is the rulebook an delegate needs to use it responsibly: how much, how soon, can I collect it, and can I alter my mind?

Having the data isn’t enough. An delegate needs information, context, rules, and operational capabilities to determine a customer’s decision. For now, most stores that appearance prepared by default are lone prepared until the buyer asks a genuine question.

One item our test did not measure, and that amazed me most in hindsight: we told all shop exactly which item to buy. A genuine buyer rarely arrives alongside a type ID. They need assistance choosing between sizes, models, and connected products, and that decision logic is merchant knowledge no phase can instal by default. Shopify can transport it; lone the merchant can compose it. That’s the next test, and the topic of a forthcoming column.

Methodology: 122 foremost Shopify storefronts, selected from published rankings of top Shopify merchants, probed between August and September 2026. The layers tested were: agents.md, llms.txt, UCP profiles and MCP tool listings; product-page organized data; the apparent content of guideline pages; the buyer questions and authority questions put to all store’s own guideline tool; a non-completing checkout lifecycle test (never calling complete_checkout or sending payment); a static-HTML accessibility inspect of all merchandise page; and, on September 30, a browser inspect of the new WebMCP checkout tools at two stores, study the checkout without submitting it. Requests were rate-limited, and responses that were throttled or blocked were recorded as unidentified fairly than absent. Coverage varies by test: guideline content for 56 stores, guideline tool for 88, live checkout for 74, product-page accessibility for 68. Shipping was tested to one Boston address. The accessibility inspect says server HTML only, so it understates what JavaScript may add.

More Resources:

• Getting Your Product Into ChatGPT Isn’t The Hard Part, Getting It Through Checkout Is
• The Web Is Growing A Second Layer – Almost A Third Head
• Does My Product Page Copy Still Matter If Agents Read Feeds And Schema? – Ask An SEO

Featured Image: Woonsen16/Shutterstock