Are Shopify stores ready for AI shopping agents? We audited ~3,700 stores (2026)

What is agentic commerce?

Agentic commerce is shopping where an AI assistant does part of the work for the buyer: it searches, compares products and, on some channels, completes the checkout inside the chat.

For a Shopify merchant, it means your products can be found and sold in ChatGPT, Google AI Mode, Gemini, Microsoft Copilot and Meta’s apps, using data Shopify sends on your behalf.

The term covers two different things, and merchants often mix them up.

1. Discovery. An assistant answers “best waterproof hiking boots under $200” with a list of products, prices and links. ChatGPT works this way for Shopify stores today.
2. Agent checkout. The assistant also takes payment, so the buyer never visits your site. Copilot, Google AI Mode, Gemini and Meta offer this for eligible Shopify stores.

Both run on structured product data. An agent doesn’t see your hero banner. It reads titles, prices, stock status, variant names, barcodes, and shipping and return terms. Holiday 2026 is the first Black Friday season with Shopify’s AI channels switched on by default.

What are Shopify Agentic Storefronts?

Agentic Storefronts is Shopify’s feature that syndicates your products to AI channels from the Shopify admin.

It sends your catalogue through Shopify Catalog (to ChatGPT and Copilot), the Google & YouTube channel (to AI Mode and Gemini) and the Facebook & Instagram channel (to Meta). It is on by default for eligible stores, with no extra app or transaction fee.

Shopify introduced Agentic Storefronts in the Winter ’26 Edition on December 10, 2025 (Shopify, Dec 10, 2025). On March 24, 2026, Shopify said products “become discoverable in ChatGPT by default via Agentic Storefronts, with no separate integrations, no apps, and no transaction fees” (Shopify, Mar 24, 2026). The Spring ’26 “Everywhere” Edition on June 17, 2026 added one Agentic area in the admin that reports orders, sales and conversions from ChatGPT, Copilot, AI Mode, Gemini and Shop (Shopify, Jun 17, 2026).

Where to manage it: Sales channels > Agentic in your Shopify admin. You must accept the Agentic Storefronts Supplemental Terms (Shopify Help, as of Oct 2026).

Which AI channels does Shopify send products to?

Sources: Shopify Help pages for ChatGPT, channels with built-in checkout, Google, Meta and Catalog requirements, all as of October 2026.

For Canadian merchants: ChatGPT and Copilot accept stores based outside the US, as long as you sell to US customers. Google AI Mode and Gemini checkout currently require a store based in the US. Meta covers Canada.

Shopify says AI searches powered by Shopify Catalog “convert at 2x the rate of those using scraped data” (Shopify, Jun 17, 2026). That is Shopify’s own figure, not an independent one.

What is the Universal Commerce Protocol (UCP)?

The Universal Commerce Protocol is an open standard that lets AI agents search a merchant’s catalogue, build a cart and check out through one common interface.

Google launched it on January 11, 2026, co-developed with Shopify, Etsy, Wayfair, Target and Walmart. Shopify stores are UCP-enabled by default, so merchants don’t need to set anything up.

Google announced UCP at NRF and listed endorsers including Adyen, American Express, Best Buy, Macy’s, Mastercard, Stripe, The Home Depot and Visa (Google, Jan 11, 2026). A March 19, 2026 update added multi-item carts, real-time catalogue data and identity linking for loyalty programs (Google, Mar 19, 2026).

The specification at ucp.dev covers catalogue search and lookup, cart building, identity linking, checkout and order management. It supports REST and JSON-RPC, with built-in support for MCP, A2A and AP2.

On a Shopify Online Store, you can see your own UCP profile at yourstore.com/.well-known/ucp. It is a JSON file that tells agents which protocol versions your store supports and where its endpoint lives. Since September 2026, store profiles declare UCP version 2026-08-25, with no merchant action needed (Shopify changelog, Sep 4, 2026). Shopify’s Spring ’26 Edition says merchants are “UCP-enabled by default” (Shopify, Jun 17, 2026).

In our audit, 99.7% of 3,703 live Shopify stores served a working UCP profile in October 2026.

Can ChatGPT buy from my Shopify store?

Not inside the chat. ChatGPT shows your products and sends the buyer to your own checkout.

OpenAI wound down Instant Checkout for merchants on March 24, 2026, and Shopify’s ChatGPT channel is now discovery-only. Buyers pay on your store, in ChatGPT’s in-app browser or a new tab. Copilot, Google AI Mode, Gemini and Meta do offer in-chat checkout.

The short history:

• September 29, 2025. OpenAI launched Instant Checkout and the Agentic Commerce Protocol (ACP) with Stripe, starting with US Etsy sellers and saying Shopify merchants were “coming soon” (Stripe, Sep 29, 2025).
• March 24, 2026. OpenAI said the first version “did not offer the level of flexibility” it wanted, let merchants use their own checkout, and refocused on product discovery (OpenAI, Mar 24, 2026; CNBC, Mar 24, 2026).
• Today. Shopify Help describes the ChatGPT channel as “a discovery-focused referrer platform”. “ChatGPT users complete their purchase on your online store checkout in a ChatGPT in-app browser, or in a new tab when customers use ChatGPT web” (Shopify Help, as of Oct 2026).

This matters for planning. ChatGPT traffic lands on your product page and checkout, so those pages still decide the sale.

On the other channels, checkout happens in the chat:

• Microsoft Copilot. Copilot Checkout launched in the US on January 8, 2026. Microsoft said Shopify merchants “will automatically be able to sell in Copilot Checkout with no application or integration needed” and would be enrolled after an opt-out window (Microsoft Advertising, Jan 2026).
• Google AI Mode and Gemini. Direct checkout is on by default for eligible stores. Subscriptions, bundles, customisable items and B2B-only products are not supported. Custom pixels and Google Analytics don’t fire on these orders; only server-side pixels do (Shopify Help, as of Oct 2026).
• Meta. Direct checkout is the default, with a redirect to your checkout when information is missing (Shopify Help, as of Oct 2026).

That pixel point is easy to miss before Black Friday: AI Mode and Gemini orders won’t show up in browser-based analytics.

How do AI shopping agents choose which products to show?

AI agents pick products from structured data: title, description, price, availability, brand, variants, identifiers like GTIN, and shipping and return terms.

OpenAI says ChatGPT uses structured metadata from first-party and third-party providers and ranks merchants on availability, price, quality and whether they are the maker or primary seller. Missing fields give an agent less to match on.

No AI company publishes a ranking formula. They do publish the data they ask for, and the lists overlap.

OpenAI’s product feed spec requires item_id, title, description, url, brand, seller_name, image_url, availability and price. It recommends color, size and material, variant grouping, gtin, mpn, shipping, return policy, reviews and extra images (OpenAI developer docs, as of Oct 2026). Shopify merchants don’t build this feed; Shopify Catalog fills it from your admin, so empty fields there stay empty in the feed.

OpenAI’s shopping help says products are chosen using “structured metadata from first-party and third-party providers (e.g., price, product description)” and that merchants are ranked by availability, price, quality and maker or primary seller (OpenAI Help).

Google’s merchant listing structured data requires a Product name, image and offers with price above zero and priceCurrency. It recommends brand, gtin, sku, aggregateRating, availability, shippingDetails and hasMerchantReturnPolicy (Google Search Central, as of Oct 2026).

Shopify’s Catalog guidance lists title, description, images, Product type, Vendor, Collections, Tags, Barcode (ISBN, UPC, GTIN), variant option names, store policies and FAQs (Shopify Help, as of Oct 2026). Its AI guide adds clear specs, sizing, material and care details, comparison information and alt text (Shopify Help).

The overlap is your to-do list: a specific title, a real description, a brand, a barcode, named variants, current stock and price, and clear shipping and return terms.

How ready are Shopify stores for AI shopping agents? (our data)

Most are eligible, few are complete.

In EVDEV’s October 2026 audit, 78.2% of 3,654 live Shopify stores passed every proven requirement we could test from outside. Gaps showed up in the product details agents and search engines recommend: GTINs, shipping and return markup, and full Product structured data.

We split the checks into proven requirements, which Shopify, Google or OpenAI state as rules for a named AI channel, and best practice, which a platform recommends but doesn’t require.

Proven requirements: pass rates

Across 509,490 products, 97.0% had a title, an image and a price above zero.

The policy check is where stores failed most often among the proven rules. 7.8% had no refund policy. That takes ten minutes to fix in Settings > Policies, and every AI channel requires it.

We couldn’t test admin-only settings: channel toggles, Catalog opt-outs, Merchant Center, plan, account standing or US selling. As a weak proxy for the last, 78.3% of stores used USD as base currency.

Which product data are Shopify stores missing?

The biggest gaps are identifiers and policy markup.

In October 2026, 35.0% of 3,659 Shopify best-seller pages had a GTIN in their structured data, 7.9% had shipping details and 8.4% had a return policy. Price, currency and availability were present on most pages that had Product markup at all.

We read every JSON-LD block on one product page per store: the store’s best seller. 87.9% had JSON-LD Product markup. Another 5.9% used older microdata only, and 6.2% had no Product markup at all.

The barcode gap

GTINs are the clearest example of data that exists in Shopify but doesn’t reach the page. 49.7% of sampled products had at least one check-digit-valid GTIN in the variant Barcode field. Yet 16.9% had a valid GTIN in Shopify that was missing from the page’s structured data.

The likely cause is the theme. Shopify’s own structured_data Liquid filter outputs brand, description, image, name, price, currency and availability, but its documentation doesn’t list gtin, shippingDetails or hasMerchantReturnPolicy (Shopify Liquid docs, as of Oct 2026). Themes that rely on it, or hand-roll a short version, leave those fields out.

Shopify Catalog reads the barcode from your admin, so the AI channels still get it. The page markup matters for Google’s merchant listings and for agents that read your site directly.

Pages vs markup: the policy gap

Many stores have written policies that machines can’t read. 82.6% of stores had a refund-policy page but no return-policy markup anywhere on the product page or homepage. 61.4% had a shipping-policy page but no shipping markup.

How complete is the average product page?

We scored each best-seller page from 0 to 10, one point for each core field: name, image, description, brand, sku, gtin, price, priceCurrency, availability and aggregateRating. The median store scored 8. 14.5% scored 10, and 12.2% scored 0 because they had no Product JSON-LD.

Catalogue-level data

Beyond the single product page, we checked the first 250 products in each store’s public catalogue.

The median store’s typical description ran 606 characters. The 300-character line is our threshold, not a platform rule.

Which Shopify themes publish the best product data?

Your theme writes most of the structured data on your product pages. The table groups stores by theme family (renamed copies of Dawn still count as Dawn), for themes with at least 30 stores in the sample.

The most common theme in the sample was Dawn, used by 11.9% of stores. 54.0% ran a Theme Store theme.

Five premium themes (Palo Alto, Impact, Focal, Enterprise and Prestige) published a GTIN on more than half of their product pages, against 35.0% across all stores. Debut, Shopify’s older free theme, had the lowest median score and the fewest GTINs in markup. No theme marked up shipping or returns on more than a quarter of its pages, so switching themes won’t close that gap on its own.

Should I block GPTBot and other AI crawlers on my Shopify store?

Blocking training bots like GPTBot doesn’t hide your products from AI shopping.

It only stops that company using your pages to train models. Search bots such as OAI-SearchBot and PerplexityBot are different: blocking them removes your pages from that engine’s answers. Neither affects the products Shopify Catalog sends to your Agentic Storefronts channels.

Shopify Help says it plainly: “Blocking AI crawlers at the /robots.txt or network layer affects only open-web discoverability. It doesn’t stop your product data from being sent by Shopify Catalog to the agentic storefronts that you’ve activated” (Shopify Help, as of Oct 2026).

So there are three separate pipes:

Sources: OpenAI bots, Perplexity bots, Google AI features, Google common crawlers (updated Jul 14, 2026), as of October 2026.

Google says Google-Extended “does not impact a site’s inclusion in Google Search nor is it used as a ranking signal.” OpenAI says its bots take about 24 hours to pick up robots.txt changes.

What Shopify stores actually do

Shopify’s current default robots.txt blocks no AI user agent, so any block is a merchant’s own change.

• 0.4% (13/3,688) of 3,688 stores blocked at least one AI search bot (OAI-SearchBot, PerplexityBot, Claude-SearchBot or Applebot) in October 2026.
• 0.8% (31/3,688) blocked at least one training bot (GPTBot, Google-Extended, ClaudeBot, Applebot-Extended or CCBot). 0.7% blocked GPTBot specifically.
• 0.1% blocked Googlebot or bingbot.
• 4.4% named any AI bot in robots.txt at all.

The robots.txt problem nobody talks about

Shopify’s current default allows /policies/ and /search and points agents to /agents.md and the UCP endpoint. Stores that once customised robots.txt.liquid often still serve an older rule set that blocks both.

84.6% of stores served Shopify’s current default unchanged. 8.7% served the older template, and 12.6% blocked /policies/ for all bots. Blocking the policy pages hides your refund and shipping terms from any crawler that reads them.

A catch for developers: in October 2026, a Shopify community thread reported that the robots.default_groups Liquid object, used in custom robots.txt.liquid files, still returns an older rule set than the default Shopify serves. Shopify staff replied they are looking into it (Shopify developer community, Oct 2026). If you have a custom robots file, check the output, not the template.

The takeaway: almost no Shopify store blocks AI search bots, but about one in eight still hides its policy pages from every crawler. If your robots.txt was customised years ago, compare it with Shopify’s current default.

Does llms.txt matter for a Shopify store?

Not much, and you already have one.

Every Shopify Online Store automatically serves /llms.txt, /llms-full.txt and /agents.md. Shopify calls /agents.md the canonical file and says the others exist for older AI crawlers. Google says Search doesn’t use llms.txt, and we found no primary documentation from OpenAI, Perplexity or Microsoft saying their shopping features read it.

Shopify Help: “Every Shopify store automatically serves the following discovery URLs … /agents.md — this is your canonical agent discovery URL … /llms.txt and /llms-full.txt — these URLs are compatible with older AI crawlers … You don’t need a third-party app to generate these files.” It also notes that Shopify Catalog, not these files, is “the authoritative product data feed to your agentic channels” (Shopify Help, as of Oct 2026).

Google’s generative AI optimization guide says: “You don’t need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search … as Google Search itself doesn’t use them” (Google Search Central, updated Jul 10, 2026).

Since May 28, 2026, merchants can edit these files with templates/agents.md.liquid, templates/llms.txt.liquid and templates/llms-full.txt.liquid in the theme (Shopify changelog, May 28, 2026).

In our audit, 95.1% of 3,703 live stores served Shopify’s default llms.txt, 4.3% served a customised one and 0.5% served none. Because Shopify generates the file, its presence tells you nothing about a merchant’s readiness, and we don’t score it.

Our advice: don’t pay for an llms.txt app. If you customise agents.md, add shipping regions, return window and sizing notes, then spend the real time on product data.

What do merchants get wrong about AI shopping?

The most common mistakes are about llms.txt, bot blocking, ChatGPT checkout and Perplexity.

Each one sends merchants to the wrong fix. Here is what the platforms’ own documentation says, as of October 2026.

What should Shopify merchants fix before Black Friday 2026?

Fix the proven requirements first, then the product data.

Make sure your store isn’t password-protected, your three policies are filled in, every product has a title, image and price, and no product you want sold is Unlisted or hidden from search. Then add barcodes, brand, product types and variant names, and check that your theme publishes them.

Black Friday falls on November 27, 2026. Most fixes below are admin work, not development.

Proven requirements (do these first)

These are stated requirements for Shopify’s AI channels (Shopify Catalog requirements, as of Oct 2026).

• Store is live, not password-protected.
• Policies filled in at Settings > Policies: Terms of service, Privacy policy, Refund policy.
• Every product has a title, at least one image and a price above $0.
• Products are published to the Online Store (or Hydrogen/Headless) channel with a working product URL.
• Products you want sold are not Unlisted and not hidden from search engines.
• You sell to US customers. Required on all channels; Google also needs a US-based store.
• Channels are on at Sales channels > Agentic, and you’ve accepted the Agentic Storefronts Supplemental Terms.
• Google: Merchant Center synced through the Google & YouTube channel. Required for AI Mode and Gemini.
• Meta: Facebook & Instagram channel with product sync on (US, Canada, Mexico).
• robots.txt doesn’t block search bots you want answers from: OAI-SearchBot, PerplexityBot, Googlebot, bingbot.
• Product pages don’t carry noindex, nosnippet or noarchive.

Best practice (do these next)

Recommended by Shopify, Google or OpenAI, but not hard rules.

• Barcode (GTIN, UPC, ISBN) on every variant. Shopify, Google and OpenAI all ask for it.
• Vendor filled in with your real brand name. OpenAI’s feed spec requires brand.
• Product type, Collections and Tags set.
• Variant options named (“Size”, “Color”), not left as “Title”.
• Descriptions with specs: size, material, care, fit, compatibility.
• Image alt text that describes the product.
• Theme outputs full Product JSON-LD: add gtin, brand, sku, shippingDetails and hasMerchantReturnPolicy. Check one product page with Google’s Rich Results Test.
• Review app outputs aggregateRating in the page HTML, not only through JavaScript.
• Install the free Shopify Knowledge Base app and add FAQs on shipping, returns and sizing. Shopify calls these FAQs “a trusted data source for AI platforms” (Shopify App Store).
• Server-side tracking for AI Mode and Gemini orders, since custom pixels and GA don’t fire there.
• If you use a custom robots.txt.liquid, check that /policies/ and /search aren’t blocked.

Optional

• Customise agents.md with shipping regions, return window and sizing notes.
• Review which products you exclude. To remove a product from ChatGPT or Copilot only, remove Catalog access for that channel (Shopify says this takes up to 7 days). To hide it everywhere, mark it Unlisted.

How can I check if my Shopify store is ready for AI agents?

You can run the basic checks yourself in about 15 minutes.

Open your robots.txt, your /.well-known/ucp file and your three policy pages, run one product page through Google’s Rich Results Test, and check that your best sellers have barcodes. For the automated version of the checks in this report, run EVDEV’s free AI-readiness checker.

Step by step:

1. Visit yourstore.com/robots.txt. Look for User-agent: OAI-SearchBot, PerplexityBot, Googlebot or bingbot followed by Disallow: /, and check that /policies/ isn’t disallowed.
2. Visit yourstore.com/.well-known/ucp. You should see JSON with a version. A 404 usually means a headless storefront.
3. Visit /policies/refund-policy, /policies/privacy-policy and /policies/terms-of-service. Each should show real text.
4. Run your best-selling product page through Google’s Rich Results Test. Check for Product with price, availability, brand and GTIN, plus shipping and return details.
5. Open your top 10 products in the admin. Check that each variant has a Barcode, the product has a Vendor and Product type, and variant options are named.
6. Open Sales channels > Agentic. Confirm which channels are on and that nothing you want sold is excluded.
7. Ask ChatGPT or Gemini for your product category. For example, “best [your product] under [price] that ships to [your country]”. Note who appears and what details they show.

How EVDEV has handled this for clients

Two recent projects show the kind of work involved. Both are factual summaries from our case studies.

• Radeckt and Lar Lar Land (United States, 2026)

  We built Shopify sites for two music acts from one management team. Part of the job was fixing the sites’ JSON-LD so each artist is indexed under the right name in Google and AI search, alongside HubSpot email capture and Google Tag Manager. Wrong entity data means an engine attaches your pages to the wrong name; the fix is in the markup.

  Read the Radeckt case study →

• VIGO (New York City)

  VIGO’s Shopify Plus B2B store carries more than 12,000 kitchen and bath products, some with over 140 metafields of specs each. EVDEV built the store with AI-powered search, advanced filters and side-by-side product comparison. A catalogue that detailed is what agents work best with, as long as the specs reach the feed and the page.

  Read the VIGO case study →

How we ran the audit (methodology)

FAQ