How to Track Sales From ChatGPT (and Other AI Assistants) in Shopify

Last checked: October 13, 2026.

Shoppers now ask ChatGPT, Copilot, Perplexity, Gemini and Claude what to buy, then click through to stores. Shopify reported that orders from AI-referred traffic grew nearly 13x year over year in Q1 2026 (Shopify Enterprise blog). So "how much is AI actually selling for me?" has become a real question. Here are four ways to answer it, from easiest to most detailed, and what each one misses.

Method 1: Shopify's built-in Agentic reports (start here)

If your store is eligible for agentic storefronts, go to Sales channels > Agentic. Shopify provides per-channel reports for sales, orders, online store sessions and online store conversion, with date filtering, plus a Total sales report across all AI channels in the Explore section. Shopify says these sales "include both referral-based sales and direct checkout sales" (Shopify Help: Managing agentic storefronts).

Covers: ChatGPT, Microsoft Copilot, Google AI Mode/Gemini and Meta surfaces, the channels connected through agentic storefronts. It also counts direct checkouts inside Copilot or Gemini, which never touch your storefront. (Shopify says GA4 and custom pixels don't fire in Copilot's direct checkout; see selling on Microsoft Copilot with Shopify.)

Misses:

• Assistants that aren't agentic-storefront channels, such as Perplexity and Claude, when they send a normal link click.
• Headless stores. Shopify says they don't get agentic performance analytics yet.
• Influence without a click. If someone learns about you in ChatGPT and later types your URL or searches your brand on Google, that sale shows up as direct or organic. (This is true of every method here.)

Method 2: Look at the order's customer journey

Open any order in your admin. The order page has a conversion or customer-journey summary showing how the customer arrived: the referring site, the landing page (including any UTM parameters), and the number of visits before purchase. An order whose referrer is chatgpt.com or whose landing page contains utm_source=chatgpt.com came through ChatGPT.

This is the most direct evidence you'll get, but it's one order at a time. Use it to check individual orders and to sanity-check whatever report you rely on.

(The exact label and layout of this section has changed over time. If you can't find it, look for "Conversion summary" on the order page.)

Method 3: Filter sessions and orders by AI referrer or UTM source

Many AI assistants add a utm_source parameter to outbound links. Commonly observed values (Highview Apps knowledge base):

In Shopify Analytics, build a sessions or sales report that breaks down by referrer source or UTM source, then filter with "contains" on those values. Highview suggests the ShopifyQL pattern utm_source CONTAINS 'chatgpt', and CONTAINS '.ai' to catch smaller assistants whose domains end in .ai. Save the report and check it monthly.

Watch out for:

• Missing referrers. Some app in-app browsers and privacy settings strip the referrer, so the visit shows as "direct". This is why the UTM parameter matters: it survives when the referrer doesn't.
• First touch vs last touch. A customer may arrive from ChatGPT on Monday and buy on Friday from a Klaviyo email. Last-click reports give the sale to email. Look at both first and last visit if your tool allows it.
• Your own links. If you share links with utm_source=chatgpt in your own content, you'll pollute the data. Keep AI UTM values for real AI referrals.

Method 4: Google Analytics 4

If you run GA4 on your store, create a custom channel group with an "AI assistants" channel matching sources like chatgpt.com, perplexity.ai, claude.ai, gemini.google.com, copilot.microsoft.com and grok.com. Place it above the default "Referral" channel so those sessions don't get lumped in with generic referrals. Then use the Traffic acquisition and Purchase reports with that channel group.

GA4 is good for trends and landing-page analysis (which product pages AI sends people to). It won't match Shopify order totals exactly, because of consent banners, ad blockers and attribution-model differences, so treat Shopify's order data as the source of truth for revenue.

What "good" looks like: three numbers to track monthly

1. AI orders and AI revenue, by assistant. Is it growing, and which assistant drives it?
2. AI share of total revenue. Shopify notes organic search still refers more sessions than all AI platforms combined (Shopify Enterprise blog), so expect a small share that's growing.
3. Landing products. Shopify found more than half of AI-referred sessions start on a product page, compared with about 20% for organic search (same source). Which product pages get AI visitors tells you which products assistants understand well. Improve the product data of the ones that should be there and aren't.

Why the numbers will always undercount

Every method above measures clicks that end in orders. AI assistants also influence purchases without a click: a shopper asks for a recommendation, remembers your brand, and buys a week later. Treat AI attribution as a lower bound and watch the trend, not the absolute number. Also ask new customers "How did you hear about us?" with a post-purchase survey, and include "ChatGPT / AI assistant" as an option.

Putting it together

Done monthly, this takes about 15 minutes, and it tells you whether improving your product data for AI channels is paying off.

Sources

• Shopify Help, Managing agentic storefronts
• Shopify Enterprise blog, "AI-referred shoppers convert better and spend more" (May 11, 2026)
• Highview Apps, AI utm_source values
• Shopify blog, Perplexity shopping (Apr 2, 2026)