How to Track ChatGPT Traffic and Other AI Referrals in GA4 and Shopify

To track ChatGPT traffic in GA4, start with the built-in AI Assistant channel in the Traffic acquisition report. Then build a custom channel group that matches AI referrers by session source with a regex, so you also catch older data and assistants Google doesn't recognize. In Shopify, open Sessions by referrer and Total sales by referrer and filter for ChatGPT and the other assistants.

Neither tool sees everything. Clicks from Google's AI Overviews and AI Mode count as Google organic, some AI clicks arrive with no referrer and land in Direct, and many buyers read an AI answer without clicking at all. Below is the exact setup we use on Shopify stores, the regex you can paste, and what to use for the parts analytics can't see.

What utm_source=chatgpt.com means#

utm_source=chatgpt.com is a tracking parameter ChatGPT adds to many of the links it shows. When someone clicks through to your store, the landing URL carries it, so analytics tools can attribute the visit to ChatGPT. You don't need to add it and you can't control it.

It has one side effect in GA4. When a URL carries UTM parameters, GA4 uses them instead of the referrer, and ChatGPT sets a source but no medium. Depending on how the click arrives, you may see the visit as chatgpt.com / referral, as chatgpt.com / (not set) inside the Unassigned channel, or, since May 2026, with the ai-assistant medium. That's why every rule below matches on source, not medium.

Step 1: Check GA4's built-in AI Assistant channel#

In May 2026 Google added an AI Assistant channel to GA4's default channel group. Its definition: the medium exactly matches ai-assistant, and GA4 sets that medium (and the campaign (ai-assistant)) when the referrer matches its list of AI assistants.

To see it, go to Reports > Acquisition > Traffic acquisition and keep the primary dimension on the default channel group. If you've had AI visits since the rollout, AI Assistant appears as its own row.

Know its limits before you report on it:

• Google hasn't published the full list of assistants it recognizes. Its help page names examples (ChatGPT, Gemini, DeepSeek, Copilot, Grok), so you can't check coverage.
• It excludes Google's AI Overviews and AI Mode. Those clicks stay in Organic Search.
• Treat it as forward-looking. Google hasn't said it relabels sessions from before the rollout. Custom channel groups, which do apply to historical data, are how you see the trend.
• It needs a referrer. Visits that arrive without one land in Direct, or in Unassigned if they carry only a utm_source.

Step 2: Build an AI channel group in GA4#

A custom channel group fixes most of those gaps. Standard GA4 properties allow two custom groups with up to 50 channels each, and custom groups apply to past data as well as new data.

Here's the regex we use for the session source condition:

^(.*\.)?(chatgpt\.com|openai\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com|copilot\.com|edgeservices\.bing\.com|deepseek\.com|grok\.com|meta\.ai|chat\.mistral\.ai|you\.com|phind\.com|poe\.com)$

Why it's written this way:

• GA4 regex is full match by default (Google's regex reference), so the pattern describes the whole source value.
• *`(.\.)? allows subdomains**, so chat.openai.com matches through openai.com and chat.deepseek.com through deepseek.com`.
• Dots are escaped, so chatgpt.com can't match a lookalike such as chatgptxcom.
• No generic words like ai or gpt on their own, which would sweep in unrelated sites.

Now build the group:

1. Go to Admin > Data display > Channel groups.
2. Open Default Channel Group and choose Copy to create new, so every standard channel carries over. Name it something like "Channels incl. AI".
3. If the copy includes an AI Assistant channel, edit it. If not, click Add new channel and name it "AI Assistants".
4. Set the condition to Source matches regex and paste the regex. Add an OR condition: Default channel group exactly matches AI Assistant, so anything Google recognizes that your regex doesn't is still included.
5. Save the channel, click Reorder, drag it to the top (above Referral and Unassigned, which would otherwise claim the traffic first) and click Apply. GA4 puts each session in the first channel it matches.
6. Save the group.

To use it, open Reports > Acquisition > Traffic acquisition, click the primary dimension dropdown and pick your new group.

Step 3: Build an AI traffic exploration#

Channel groups tell you how much. An exploration tells you where AI visitors land and whether they buy.

1. Go to Explore > Blank (a free-form exploration).
2. Add dimensions: Session source, Session source / medium, Landing page + query string, Session default channel group, your custom channel group and Device category.
3. Add metrics: Sessions, Engaged sessions, Engagement rate, Key events, Ecommerce purchases and Purchase revenue.
4. Tab 1, sources: rows = Session source, filter Session source matches regex (paste the regex). This is your AI traffic by assistant.
5. Tab 2, landing pages: rows = Landing page + query string, same filter. Expect many visits to land straight on product pages, because assistants deep-link to the products they recommend.
6. Tab 3, comparison: rows = your custom channel group, same metrics, no filter. Compare AI visitors with Organic Search on engagement, purchase rate and revenue per session.

Before you trust the regex, run a discovery pass: in Tab 1, change the filter to Session source partially matches regex gpt|openai|perplexity|claude|gemini|copilot|deepseek|grok|mistral. Any AI source that shows up there but isn't in your channel regex gets added to it.

Purchase data only appears if GA4 receives your Shopify purchase events, which is the case when GA4 is connected through Shopify's Google & YouTube app. If purchases are missing, fix that before judging AI revenue.

Step 4: Track ChatGPT traffic in Shopify#

Shopify's reports count sessions and orders independently of GA4, and every order is recorded whether or not the visitor accepted analytics cookies, so use Shopify as the revenue reference.

1. Sessions by referrer. In the admin, go to Analytics > Reports, choose the Acquisition category and open Sessions by referrer. Shopify's own guidance is to filter by referrer channel for ChatGPT, Perplexity, Copilot or Claude, and to filter the referrer host for gemini.google.com to find Gemini.
2. UTM source. In the same report, add a filter on UTM source for chatgpt.com to catch ChatGPT clicks that carried the parameter.
3. Total sales by referrer. In the Sales category, Total sales by referrer shows revenue by referrer source and host. Filter the host for the AI domains.
4. Order conversion summary. Open any order under Orders and look at the Conversion summary. It shows where the customer's first visit came from, and View conversion details shows the visit history. For high-ticket stores, check your largest orders each month: a first visit from ChatGPT three weeks before a $4,000 order is exactly the pattern totals hide.
5. Agentic checkout. If you sell through Shopify's agentic storefronts, orders from AI channels show channel or referrer attribution in the admin. Check them in Total sales by sales channel as well. Our explainer on ChatGPT Shopping for Shopify stores covers how those orders flow.

Shopify's own data gives a sense of scale: in Q1 2026, AI referral sessions to Shopify storefronts grew more than 8x and AI-referred orders nearly 13x year over year. Small numbers today can move fast.

What GA4 and Shopify can't see#

Visibility is the part most stores never measure. Our AI visibility audit starts there: which prompts name you, which sources the assistants cite, and which competitors appear instead.

Two proxies deserve more weight than they get. A post-purchase survey ("How did you first hear about us?") with ChatGPT and "another AI assistant" as options catches visits analytics lost, and for high-ticket brands the same question belongs on sales calls. And branded search is where AI influence often shows first, which is why our guide to how AI Overviews cite stores tracks it alongside citations.

A monthly AI traffic report#

We put these in the same monthly report as rankings, traffic and revenue; the sample monthly SEO report shows the format. When the landing page tab shows which products AI assistants send people to, improve those pages first, then work on being named more often: our guides to ranking in ChatGPT and getting into the “best of” lists AI cites cover that side.