How to Track AI Traffic: The One AI Channel With a Verified Revenue Signal

TL;DR

• Tracking AI traffic takes 4 layers: GA4’s native AI Assistant channel, a custom channel group with a referrer regex, a self-reported attribution field, and visibility tracking upstream of the click.
• The reason it’s worth the setup: in NP Digital’s late-2025 survey of 1,903 marketers, 89% of those seeing revenue from AI credited traffic arriving from AI platforms. AI copy and AI ads barely registered.
• AI referral traffic converts at up to 4.4x the rate of organic search, on volumes still around 1% of total traffic for most industries.
• Default analytics hide most of it. Loamly found 70.6% of identifiable AI traffic lands in GA4’s “Direct” bucket, and Google AI Overviews pass no distinct referral signal at all.
• Whatever your dashboard shows for this channel today, treat it as a floor.

If you want revenue from AI, the data points at one use case: visitors arriving from ChatGPT, Perplexity, Gemini, and other AI platforms. Tracking that traffic takes deliberate setup, because GA4’s defaults scatter it across Direct, Referral, and Unassigned. 

The revenue case: what 1,903 marketers reported

In October 2025, NP Digital asked 1,903 marketers one question: has AI helped you generate more revenue?

91% said no.

The 9% who said yes are where it gets interesting. Asked where the revenue came from, 89% pointed to traffic from LLMs. 4% credited better conversion from AI creatives. 2% credited AI-generated content.

Across everything companies have thrown AI at over 3 years, one use case carries a clear revenue signal: AI traffic. And most businesses still miss it, because if they didn’t, the 9% would be much larger.

Where the AI budget went instead

Most AI spend chased production: copy, creatives, automated workflows. The efficiency gains are real. The revenue data is brutal.

MIT’s NANDA initiative reviewed enterprise AI in 2025: despite $30 to 40 billion invested in generative AI, 95% of pilots showed no measurable P&L impact. S&P Global Market Intelligence surveyed over 1,000 enterprises the same year and found 42% had abandoned most of their AI initiatives, up from 17% a year earlier.

I see the same pattern with clients. AI helps their teams produce content faster, and 6 months later the pipeline looks identical. Their constraint was always distribution: getting content in front of buyers on Google and inside AI answers. Faster production leaves that constraint untouched.

Why AI referral traffic converts

Someone asking ChatGPT for “the best AI visibility tool for a mid-size SaaS” has already framed their problem. They compare options inside the conversation, narrow the list, and click through with most of the research done. They arrive pre-qualified.

The conversion data agrees:

• Semrush measured AI-referred visitors converting at 4.4x the rate of organic search visitors.
• Ahrefs analyzed its own traffic in June 2025: AI platforms drove 0.5% of visits and 12.1% of signups.
• Seer Interactive’s B2B client analysis found signup conversion at 15.9% for ChatGPT traffic and 10.5% for Perplexity, against 1.76% for Google organic.
• In ecommerce, a 12-month analysis of 94 stores found ChatGPT traffic converting 31% higher than non-branded organic search.

One caveat before the setup: volume stays small, around 1% of total website traffic for most industries. Anyone selling AI search traffic as a volume channel today is overselling it. The conversion rates and the growth curve are the case, and I’d rather be positioned while competitors ignore it.

How to track AI traffic in GA4

GA4 shipped a native “AI Assistant” channel in May 2026. It automatically classifies traffic from ChatGPT, Gemini, DeepSeek, Copilot, and Grok. Treat its reporting with a grain of salt, though: Google’s own channel definition skips Perplexity and excludes AI Overviews and AI Mode, and default channel groups can’t be edited to fix that.

So the custom channel group is still the core setup:

1. In GA4, open Admin, then Data display, then Channel groups.
2. Create a new channel group and add a channel named “AI Traffic.”
3. Set the condition to Source matches regex: chatgpt\.com|perplexity\.ai|claude\.ai|gemini\.google\.com|copilot\.microsoft\.com
4. Drag the AI Traffic channel above Referral in the rule order, so it evaluates first, and save.

From then on, AI traffic shows up as its own row in your acquisition reports. Extend the regex as new platforms appear in your referral report.

Each layer of tracking catches something the others miss:

Tracking layer	What it catches	What it misses
GA4 native AI Assistant channel	ChatGPT, Gemini, DeepSeek, Copilot, Grok referrals	Perplexity, AI Overviews, sessions without a referrer
Custom channel group + regex	Any AI domain you add, including the platform GA4’s default misses	Clicks that strip referrer data, AI Overviews
Self-reported attribution field	Discovery that never produced a click	Anyone who skips the form field
Branded search monitoring	The AI-to-Google indirect path, in aggregate	Attribution to a specific platform or prompt
AI visibility tracking	Where you appear in answers, before any click	Nothing downstream: it’s the upstream layer

The AI traffic your analytics never see

A typical AI-influenced purchase: a user discovers your brand in a ChatGPT answer, clicks nothing, opens Google, types your brand name, and converts. Your analytics credit organic search. The AI assistant did the selling, Google got the credit.

This happens at scale. Loamly found 70.6% of identifiable AI traffic classified as “Direct” in GA4. Workshop Digital analyzed 181.6 million GA4 sessions and found 22% of ChatGPT-referred sessions marked “(not set)”, and 32% for Perplexity. 

AI Overviews clicks disappear into your regular organic numbers entirely.

Which means NP Digital’s 9% is a floor. Marketers can only report the revenue they can see, and most AI-driven revenue is invisible by default.

Two moves recover part of the hidden layer, on top of the channel group:

1. Add “ChatGPT / AI assistant” as an option in your “how did you hear about us?” field. Self-reported attribution catches what referrers strip out.
2. Watch branded search volume next to your AI visibility. When AI mentions rise and branded searches rise with them, that’s the indirect path surfacing.

If you only count clicks from chatgpt.com, you’re measuring the smallest slice of the channel.

Tracking upstream: AI visibility before the click

The tools above measure AI traffic after it reaches your site. The layer that predicts it sits earlier: whether your brand appears in AI answers at all, for which prompts, against which competitors.

That’s the gap a GEO strategy (Generative Engine Optimization) closes, and in practice it covers 4 things:

• Knowing which prompts matter in your category.
• Tracking where your brand shows up across ChatGPT, Perplexity, Gemini, and AI Overviews.
• Building the on-site content and off-site presence that gets you into those answers.
• Measuring AI visibility as its own KPI, separate from SEO.

At Crescendo this now sits alongside SEO in every visibility strategy we build. It’s also why we built Clairon, and we built it to cover both ends of the funnel. It tracks AI visibility at the prompt level: whether a brand appears in answers, by platform and country, against named competitors, with the sources driving those appearances. 

And it tracks AI traffic itself: connect your GA4 and it retrieves the traffic LLMs send you, broken down per AI engine, with its evolution over time and the pages AI visitors land on most.

The two layers answer different questions. Visibility tells you where the traffic will come from. Traffic tells you what that visibility earned. Reading them side by side is what turns “AI” from a line your analytics hide into an acquisition source you can manage.

FAQ

1. How do you track AI traffic in GA4?
   

   Combine GA4’s native AI Assistant channel with a custom channel group: add a channel with Source matching a regex of AI referrer domains (chatgpt.com, perplexity.ai, claude.ai, gemini.google.com, copilot.microsoft.com) and move it above Referral in the rule order. The custom group covers Perplexity and other platforms the native channel skips.

2. Does GA4 track AI traffic by default?
   

   Partially. The native AI Assistant channel (May 2026) classifies ChatGPT, Gemini, DeepSeek, Copilot, and Grok, and it skips Perplexity and AI Overviews. Sessions arriving without a referrer still land in Direct: Loamly measured 70.6% of identifiable AI traffic there.

3. What is AI referral traffic?
   

   Visitors who reach your site from an AI platform, either by clicking a link inside an answer or by searching your brand on Google after an AI conversation. The second path is larger and mostly invisible in analytics.

4. Why does AI traffic convert better than organic search?
   

   Users do the comparison work inside the AI conversation and arrive late in the decision process. Semrush measured a 4.4x conversion advantage over organic search; Seer Interactive found 15.9% signup conversion for ChatGPT traffic against 1.76% for Google organic.

5. How much website traffic comes from AI platforms today?
   

   Around 1% of total traffic for most industries, growing several hundred percent year over year. The case for tracking it rests on conversion quality and growth, and on the fact that your dashboard undercounts whatever the real number is.