What Shopify's own numbers say about AI shoppers (and what they don't)

Shopify published two data posts this year on how shoppers who arrive from AI assistants behave compared with shoppers from organic search. They are the best first-party numbers available for Shopify merchants, and they are also easy to over-read. This article sets out what the data says, where it is thin, and what a mid-sized brand should do with it. For how the engines pick products in the first place, see how AI decides which products to recommend.

What Shopify measured

The first post, published in May 2026, drew on Q1 2026 commerce data across Shopify storefronts.^[2] It reported three things about AI-referred sessions compared with organic search: more than half started on a product page versus 20% for organic, conversion rates were nearly 50% higher, and average order values were 14% higher.^[2]

The second post, published in August 2026, used Q2 data and added the growth and category findings.^[1] AI-referred sessions grew 197% year over year; organic grew 12%. When AI-referred shoppers reached a product page they converted about 80% better than organic-referred shoppers, and in spec-led categories, where buyers compare specifications, compatibility and reviews, the rate was roughly double.^[1] In taste-led categories the conversion gap was smaller, but AI introduced net-new customers at about 1.3x the rate of organic search.^[1]

The two conversion figures are not a contradiction

Readers trip on "50% higher" in one post and "80% better" in the next. They measure different things. The Q1 figure compares all AI-referred sessions with all organic sessions.^[2] The Q2 figure compares only shoppers who reached a product page.^[1] Since AI sends a far larger share of its visitors straight to product pages, the second cut will always look stronger. Quote whichever you use with its definition attached.

What the data cannot tell you

Three limits matter before you build a plan on these numbers.

• It is one platform's aggregate. Shopify did not publish absolute session or order counts, so the base behind each percentage is not public, and a 197% rise on a small base is still a small number for most stores. Shopify itself says organic remains the dominant discovery channel.^[2]
• The AI platforms are not named. The posts say "AI search" and "AI platforms" without listing which referrers were counted.^[1][2]
• Your analytics will undercount it. AI referrals without a referrer header land in Direct, and clicks from Google's AI Overviews arrive as ordinary organic. Shopify's figures come from its own platform data, not from a merchant's GA4. Expect your own report to show less than the trend suggests; here is how to read it.

Industry commentary has made the same point: Shopify's structured-data result is a platform-specific finding from one provider, useful as a signal rather than a benchmark.^[3]

Discovery moved into the chat; the sale did not

One more piece of context changes how to act on this. In September 2025 OpenAI launched Instant Checkout, with Shopify merchants slated to sell directly inside ChatGPT. In March 2026 Shopify told Digital Commerce 360 that buyers would instead discover products in ChatGPT, surfaced through Shopify's Global Catalog, and complete the purchase on the merchant's own store, in an in-app browser or a separate tab.^[4]

Put that next to the behaviour data. The AI does the shortlisting; your product page closes the sale. Half of those visitors have never seen your homepage.^[1] The page has to confirm the research the shopper already did in the chat, or the visit ends there.

What a $1M to $25M brand should do with this

The data points to a short list, and it is the same list that helps organic search. Google has said its AI features run on the same core ranking systems as Search, and Shopify makes the same argument: the product data that makes a catalog legible to AI is the data that gets products found in organic results.^[1]

1. Start where the research happens

The conversion advantage concentrates in spec-led categories: anything shoppers compare on specifications, compatibility, materials or use case before buying.^[1] A t-shirt and a watch sit in the same apparel category and need different levels of research.^[1] Pick the twenty to fifty SKUs in your catalog where customers ask the most pre-purchase questions and start there, not with the whole store.

2. Fix the structured product data before the copy

The 2x result for structured Catalog data is the strongest argument in either post, and it is about fields, not prose.^[1] Complete titles, product type, variants, materials, dimensions, compatibility, identifiers, price, availability, shipping and return terms, consistently across the storefront and the catalog. An AI that reads a thin record either skips the product or describes it wrongly. This is the layer our GEO work operates on, SKU by SKU.

3. Make the product page stand alone

Shopify's own guidance for deeper-funnel arrivals is a checklist: clear description, specs, use cases, compatibility, reviews, delivery and returns, FAQs and comparison-friendly detail.^[1] Put the answer to the shopper's likely question in the first screen. A visitor who arrives from a chat with a specific requirement should see it confirmed without scrolling.

4. Collect and mark up reviews

Shopify lists reviews among the most useful content a shopper reads before buying and recommends marking them up with review structured data so ratings are eligible for review snippets in Google Search.^[1] Reviews also give AI systems third-party evidence that a product does what the page says.

5. Measure before and after, on a locked set of questions

Referral reports alone will not show you whether the work changed what the AI says. Lock a set of real customer questions, run them across ChatGPT, Perplexity, Gemini and Google AI Mode on a schedule, and log whether your products are named. Change the product data, keep the questions fixed, and compare. That is the only test that attributes a visibility change to the work you did.