How to Track AI Search Traffic in Google Analytics

To track AI search traffic in GA4, filter the Session source dimension against the AI referrer hosts: chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com. The cleanest way is a custom channel group with one regex. GA4 shipped a native AI Assistant channel in May 2026, but it misses Perplexity and every visit that arrives without a referrer, so you still want your own filter.

That is the whole play. The rest of this post covers the setup in detail, with the exact hosts and regex, and explains why the number you end up with is a floor rather than a full count.

Why AI traffic is invisible in GA4 by default

By default, GA4 files AI assistant visits under Referral or Direct, mixed in with every other site that links to you. There is no “AI” row waiting in your reports. A click from ChatGPT looks like a referral from chatgpt.com, and a click from the ChatGPT app usually looks like nothing at all: a Direct session with no source. If you never go looking, you will never see it.

The Direct problem is the bigger one. Clickport documents four separate mechanisms that strip the referrer before a ChatGPT visitor reaches you: ChatGPT’s referrer policy, rel="noreferrer" on some link types, the iOS and Android apps opening links in wrappers that drop the header, and plain copy-and-paste, which never carries a referrer. Each of those sessions lands in Direct and looks identical to someone typing your URL from memory.

Put plainly, GA4 can only show you the AI visits that announce themselves. Your job is to stop those from drowning in the Referral bucket.

The referrer hosts to look for

Five hosts cover most identifiable AI search traffic today: chatgpt.com and chat.openai.com (ChatGPT), perplexity.ai (Perplexity), gemini.google.com (Gemini), copilot.microsoft.com (Copilot), and claude.ai (Claude). Match them in the Session source dimension. Add deepseek.com, grok.com, and meta.ai if your audience uses those assistants.

One quirk worth checking in your own reports: some sites see app-style sources such as com.openai.chat. That is the ChatGPT iOS app’s bundle identifier leaking through as a referrer. It will not match a regex built only for chatgpt.com, so if you spot any source starting with com.openai, count it as ChatGPT and widen your pattern.

Note that this list is about people clicking through from AI answers. The bots that read your site to build those answers are a different topic; we cover them in our guide to AI crawlers.

The native AI Assistant channel helps, but it is not enough

In May 2026, Google added an AI Assistant channel to GA4’s Default Channel Group. When GA4 recognizes an AI referrer, it sets the medium to ai-assistant and groups the session under the new channel automatically. No setup needed. Google named ChatGPT, Gemini, and Claude as examples, but has not published the full list of recognized referrers.

Two gaps make it insufficient on its own. First, Search Engine Journal found that Perplexity is not on Google’s list, so those sessions keep landing in Referral. Second, the same analysis showed the source chatgpt.com sitting in three different channels at once: AI Assistant, Referral, and Unassigned. If you read only the AI Assistant row, you miss the other two slices of the same traffic.

The fix is to define the channel yourself.

Build a custom AI channel group, step by step

A custom channel group takes about five minutes and reclassifies AI sessions in your standard reports going forward. Analytics Mania’s walkthrough is the reference here. The steps:

1. Go to Admin, then Data display, then Channel groups. Create a new channel group (do not edit the default one).
2. Click Add new channel and name it something like “AI traffic”.
3. Set the condition to Session source matches regex and paste a pattern like:

chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com|claude\.ai|deepseek\.com|grok\.com|meta\.ai

4. Drag your AI channel above the Referral channel in the list. Order matters: GA4 assigns each session to the first channel it matches, so if Referral sits higher, it swallows your AI traffic.
5. Save. Then open Reports, Traffic acquisition, and switch the primary dimension to your new channel group.

One caveat: channel groups are not retroactive. They classify data from the day you create them. For history, use an exploration.

Build an exploration to see what AI visitors do

An exploration answers the questions a channel row cannot: which pages do AI assistants send people to, and do those people do anything. Orbit Media’s guide shows the pattern. Create a blank exploration, add Session source and Landing page as dimensions, add Sessions and Key events as metrics, then filter Session source with the same regex as above.

The landing page column is the interesting part. It tells you which of your pages AI engines actually cite and link, which is direct feedback on your generative engine optimization work. If a page you never promoted keeps pulling ChatGPT visitors, that page has language the models like. Study it and repeat it.

Add a Key events column and you can compute conversion rate per AI source right in the table. That rate is the reason this whole setup is worth the effort.

AI visitors convert better, which is why this matters

The volume of AI referral traffic is still small for most sites. The value per visitor is not. A Semrush study reported by PPC Land found AI search visitors worth 4.4 times more than traditional organic visitors, measured by conversion rate. The proposed reason is simple: the assistant already answered the visitor’s basic questions, so the person who clicks through has done their early research and is closer to a decision.

Retail data points the same way. Adobe’s Q1 2026 figures, drawn from over one trillion visits to large North American retailers, showed AI referral traffic converting 42% better than non-AI traffic in March 2026. A year earlier the same channel converted 38% worse, so the quality of AI-referred visitors flipped hard in twelve months. Worth noting: Adobe published this alongside its own AI optimization product, so treat the exact number as vendor-stated and directional.

Directional is enough. If a channel converts multiples better than organic, you want it in its own row in your reports, not smeared across Referral and Direct where nobody will ever notice it. The same logic drives answer engine optimization: the clicks are fewer, but each one is worth more.

Tag the links you place yourself

Everything above covers links the AI engines create. For links you control, do not rely on referrers at all: add UTM parameters. If you publish content on other platforms, drop links in communities, or run outreach that AI assistants might surface, tag those URLs with utm_source and utm_medium so attribution survives even when the referrer does not.

ChatGPT itself does some of this for you. Citation links in its search-style answers often carry utm_source=chatgpt.com, which GA4 reads even when the referrer is stripped. In Clickport’s measured traffic, over a third of identified AI visits arrived with a UTM tag and no referrer at all. Those sessions are only countable because of the tag. Build your channel regex to match on source, and it will catch UTM-tagged sessions and referrer sessions alike.

Remember: this is a floor, not a full count

However well you configure GA4, a large share of AI-assisted visits will never identify themselves. A Loamly analysis of 20,428 visits in February 2026 found that 70.6% of AI visits arrived without referrer headers. Add the people who read an AI answer about you and then search your brand name, or type your URL directly, and the invisible share grows further. No analytics setup recovers those sessions.

So report the number for what it is: a verified minimum. If your AI channel shows 200 sessions, the true count of AI-influenced visits is meaningfully higher. The floor is still useful. The trend is real, and it shows you which engines and which pages are working. It also gives you a conversion rate for the visitors you can see.

One habit closes part of the gap: watch branded search and Direct traffic alongside the AI channel. A rise in all three at once usually means the assistants are talking about you more than your referrer data admits.

We track this loop ourselves. Gantra runs 12 marketing agents for startups, and its daily SEO and GEO audits, which are on the free plan, watch how your pages perform in both classic search and AI answers. The article agent writes content shaped for citation, and publishing runs on autopilot for LinkedIn and X, with one-click drafts for Reddit, Hacker News, Medium, and Dev.to. GA4 tells you what AI traffic arrived; the audits help you earn more of it, as part of a full AI marketing team.

The takeaway

AI search traffic is already in your GA4 property, hiding in Referral and Direct. Build a custom channel group that matches Session source against the AI hosts, keep it above Referral, and back it with an exploration that shows landing pages and conversions. Trust GA4’s native AI Assistant channel as a starting point, not a total, because it misses Perplexity and every referrer-less visit. Tag your own links with UTMs. Then read the number as a floor and watch its trend, because the data so far says these are some of the best visitors you get. If you want agents doing the earning side while you watch the numbers, Gantra starts free.