Summary

AI search tools are now sending real visitors to South African websites, but Google Analytics 4 buries that traffic, often lumping it under 'Referral' or even 'Direct'. If you cannot see it, you cannot prove it is growing or optimise for it. This guide shows you exactly how to isolate AI referral traffic in GA4 using explorations, custom segments and a custom channel group, which AI domains to include, and how to interpret what you find. It is one of the most practical things a business can do right now to understand the shift from classic search to AI answers.

Why AI traffic is hard to see in GA4

When someone clicks a link inside a ChatGPT answer, a Perplexity citation or a Gemini response and lands on your site, GA4 records the referrer, but it does not have a built-in 'AI' channel. Depending on the tool and how the link is constructed, that visit often falls into the generic 'Referral' channel, or sometimes 'Direct' when no referrer is passed. The result is that AI-driven traffic is real but invisible unless you go looking for it.

This matters because AI search is growing fast. If you cannot measure it, you cannot tell whether your AEO and GEO efforts are working, and you cannot make the case for investing in them.

Which AI domains to track

Build your filters around the domains that AI tools use when they refer traffic. The current key ones include:

• chatgpt.com and chat.openai.com (ChatGPT)
• perplexity.ai (Perplexity)
• gemini.google.com (Google Gemini)
• copilot.microsoft.com and bing.com (Microsoft Copilot)
• claude.ai (Claude)

Keep this list updated, as new tools and domains appear regularly. The point is to capture the cluster of AI assistants as a group so you can see the trend.

Method 1: A custom exploration with a regex filter

The fastest way to see AI traffic is an exploration:

• In GA4, go to Explore and create a blank exploration.
• Add 'Session source / medium' and 'Session source' as dimensions, and 'Sessions', 'Engaged sessions' and 'Conversions' as metrics.
• Add a filter on Session source using a 'matches regex' condition listing the AI domains, for example: chatgpt|openai|perplexity|gemini|copilot|claude.
• Apply it and you will see, in one view, how much traffic each AI tool is sending and how it behaves.
• Save the exploration so you can revisit it monthly.

Why regex: A single regex filter future-proofs your view: you add new AI domains to one expression rather than rebuilding the report.

Method 2: A custom channel group (the durable solution)

For ongoing reporting, create a custom channel group so 'AI Search' shows up alongside Organic, Paid and Direct in your standard reports:

• In Admin, under Data display, open Channel groups and create a new custom channel group.
• Add a new channel named 'AI Search' and order it near the top so it captures matching sessions first.
• Define the rule as Session source matches regex for your AI domains.
• Save. New data will be classified into the AI Search channel going forward.

Note that custom channel groups apply going forward and to a limited window of historical data, so set this up sooner rather than later to start building a clean trend.

What to do with the data

Once you can see AI traffic, watch three things: the trend (is it growing month on month), the behaviour (do AI-referred visitors engage and convert, or bounce), and the landing pages (which of your pages AI tools actually cite and send people to). Those cited pages are gold: they tell you what content AI engines trust, so you can create more like it.

If AI referral traffic is climbing, that is your signal to invest more in answer-first content, structured data and the kind of clear, citable writing that AI assistants surface. Measurement turns AI search from a buzzword into a managed channel.

Related Juicy Designs resources

• Analytics and reporting services
• AI SEO: SEO, AEO and GEO combined
• AI search visibility audit
• Consent Mode v2 and GTM for SA sites

Frequently asked questions

Can GA4 track traffic from ChatGPT?

Yes, but not by default. GA4 records the referrer when someone clicks a link inside ChatGPT, but it does not group it as an AI channel. You need a custom segment, exploration or channel group filtering for AI domains to isolate and measure it.

Why does AI traffic show up as Direct in GA4?

Some AI tools do not pass a referrer when a user clicks a link, so GA4 has no source to attribute and files the visit under Direct. Where a referrer is passed, it usually lands in the Referral channel until you create a custom AI channel.

Which AI domains should I track in GA4?

Start with chatgpt.com, chat.openai.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and claude.ai. Use a regex filter so you can add new AI domains to a single expression as the landscape changes.

How do I create a custom channel group for AI search?

In GA4 Admin, under Data display, open Channel groups, create a custom group, add a channel called 'AI Search', and set its rule to Session source matching a regex of AI domains. It classifies traffic going forward.

Is AI referral traffic worth measuring for a small business?

Yes. Even modest AI traffic is a leading indicator of where search is heading. Measuring it lets you see which pages AI tools cite, prove whether your AEO and GEO work is paying off, and decide where to invest.

Does tracking AI traffic require extra tags or code?

No extra tags are needed for basic referral tracking; it relies on the referrer GA4 already captures. You only configure segments, explorations or a custom channel group inside GA4 itself.