Tracking AI Traffic With the GA4 AI Assistant Channel

On 13 May 2026, Google added a native AI Assistant channel to the Default Channel Group inside GA4, automatically separating chatbot-driven visits from generic referral traffic.

The update matters because AI-sourced traffic is no longer a rounding error in most reporting.

Adobe Digital Insights recorded a 393 per cent year-on-year increase in AI-referred visits to United States retail sites in the first quarter of 2026.

That traffic also converted 42 per cent better than non-AI traffic by March. Only twelve months earlier, it had converted 38 percent worse.

The GA4 AI Assistant channel gives business owners and marketing teams a native way to measure this shift.

However, the channel has clear limitations that every property owner should understand before reporting on it.

What Is the GA4 AI Assistant Channel and How Does It Work

The GA4 AI Assistant channel is a new entry inside the platform’s Default Channel Group, sitting alongside Organic Search, Paid Search, Direct, and Referral.

Google confirmed the update in its Analytics Help Centre “What’s New” documentation on 13 May 2026, and broad availability across properties followed by early June.

The channel requires no manual configuration from property owners.

What the ‘ai-assistant’ Medium Means

When GA4 detects a referrer that matches a recognised AI assistant, it automatically assigns the medium value ai-assistant to that session.

This single dimension change triggers the rest of the classification, so property owners do not need to build custom regex rules or filters to see it appear in standard reports.

How GA4 Assigns the Channel

Once the medium is set, GA4 groups the session under the AI Assistant channel and stamps the campaign as (ai-assistant), the same convention Google uses for values like (direct) or (not set).

AI Assistant vs Referral Traffic

Before this update, a visit from ChatGPT or Gemini looked identical to a link from a news site or a partner blog. Both landed in Referral.

The AI Assistant channel now separates recognised AI sources into their own bucket. Any platform absent from Google’s recognised list, including Perplexity, still falls into Referral rather than the new channel.

Which AI Tools Does the GA4 AI Assistant Channel Track

Google recognises a specific set of AI platforms for this channel, and that list has already changed once since launch.

AI Platforms Currently Recognised by GA4

Google’s launch announcement on 13 May 2026 named three AI platforms. By June, the live Default Channel Group documentation named five instead.

Claude no longer appears in the published definition, so confirm the current list against Google’s help page before quoting it in a report.

Why Claude and Perplexity Need Separate Consideration

Claude’s absence from the current published definition does not necessarily mean GA4 has stopped recognising it.

Property owners should not assume its traffic is captured until Google confirms this directly.

Perplexity presents a different situation. It has never appeared in Google’s official AI Assistant definition.

Its sessions still land in Referral rather than the native channel, even though Perplexity typically passes clean referrer data.

What Does the GA4 AI Assistant Channel Track

For the platforms it recognises, GA4 provides genuine visibility into how AI-driven visits behave across a website.

1. AI Referral Sessions and Users. The channel reports session and user counts for every visit GA4 attributes to a recognised AI assistant. This gives a baseline volume metric inside Traffic Acquisition reports, with no custom report required.

2. Engagement and Conversions from AI Traffic. Standard engagement metrics are available for the AI Assistant channel the same way they are for any other channel. This includes engagement rate, average session duration, and conversion events.

3. Landing Pages Receiving AI Traffic. Cross-referencing the AI Assistant channel against a landing page report shows which pages AI tools are sending visitors to. This helps identify the content earning citations and clicks from generative AI answers.

4. Highest-Value AI Traffic Sources. Within the channel, the source dimension still separates individual platforms. A property owner can see whether ChatGPT, Gemini, or another recognised assistant sends the highest-value traffic, rather than treating the channel as a single undifferentiated bucket.

What Does the GA4 AI Assistant Channel Miss

The channel is a floor for AI traffic, not a complete picture. Treating it as the full measurement of AI-driven visibility risks under-reporting the impact AI search is already having on a business.

AI Mentions Without Website Visits

GA4 only measures clicks. An AI assistant might answer a question using information from a website without linking to it. A user might also read the answer and never click through.

Either way, that mention produces no session and no data point inside GA4.

AI Citations That Generate No Clicks

Even when an AI assistant does cite a website, many users treat the citation as sufficient and never visit the page directly.

This mirrors the zero-click pattern already well documented in traditional Google search results, where users increasingly get their answer without leaving the results page.

AI Visibility Outside Website Analytics

Measuring whether a business is being recommended by AI tools at all sits outside GA4’s scope entirely.

This requires monitoring the AI platforms directly, a separate discipline from standard analytics reporting.

No Historical Backfill

The classification is forward-only. GA4 does not reclassify sessions that occurred before the channel was enabled on a property.

Historical AI traffic stays recorded as Referral, Direct, or whatever custom grouping was already in place.

Why AI Traffic Can Still Appear as Direct in GA4

A large share of genuine AI-referred traffic never reaches the AI Assistant channel, because classification depends entirely on a clean referrer header.

• Referrer information drives GA4 attribution. GA4 can only assign the ai-assistant medium when the referring platform passes referrer data with the click.

Industry estimates suggest 60 to 70 percent of actual AI-originated sessions arrive without this data and land in Direct instead.

• Mobile apps often strip that data. Native AI apps on iOS and Android often strip referrer data before opening a link in the device browser.

Copying a link from a chat and pasting it into a new tab has the same effect.

• Unexpected Direct traffic is worth auditing. Check landing pages and session timing within the Direct channel for patterns consistent with AI referral behaviour, such as a traffic spike after a page is likely cited in an AI answer.

This will not give exact figures, but it treats the AI Assistant channel as a floor estimate rather than a ceiling.

AI Overviews Sit Outside the AI Assistant Channel

Google’s own AI features are explicitly excluded from the AI Assistant channel definition, which surprises many business owners who assume all AI-driven visibility would be grouped together.

How GA4 Classifies AI Overview Traffic

Clicks generated through Google’s AI Overviews are recorded as Organic Search, the same channel that has always captured standard Google search results.

There is no separate line item distinguishing an AI Overview click from a traditional blue-link click inside GA4’s standard reports.

How AI Mode Fits Into GA4 Reporting

The same treatment applies to AI Mode, Google’s more conversational search experience.

Traffic from AI Mode is also folded into Organic Search rather than appearing in the AI Assistant channel or any dedicated grouping.

Report Organic Search and AI Assistants Separately

AI Overviews likely represent the highest-volume source of AI-influenced traffic for many websites, and that volume stays invisible as “AI” inside GA4.

Pairing Search Console data with GA4 becomes necessary for an accurate picture.

Combining both sources is standard practice in the technical audits we run for clients working through AI-driven traffic changes.

How to Measure AI Traffic Accurately in GA4

A single-channel view of AI traffic understates its actual impact on a business, so a layered measurement approach is worth building now.

• Use the native AI Assistants channel for baseline reporting. Use the native AI Assistant channel as the starting point. Annotate the 13 May 2026 launch date inside GA4, and caption every report noting the figure is a floor, not a complete count.

• Maintain broader tracking for additional AI sources. Perplexity and any platform not yet on Google’s recognised list will not appear in the native channel, so a supplementary tracking method remains necessary.

• Audit source and landing-page data. Cross-reference source, landing page, and Direct traffic data to catch patterns the native channel misses, particularly around mobile referrer stripping and unrecognised platforms.

• Combine GA4 with Search Console and AI visibility data. Pair GA4 with Search Console for the Organic-routed AI Overview traffic, and consider AI visibility monitoring for citations that never produce a click.

Businesses working through poor data quality in their reporting stack often find AI traffic measurement surfaces the same issues.

Is Using Custom AI Traffic Tracking Still Important?

The native channel reduces, but does not remove, the case for custom tracking.

When the Native GA4 Channel Is Enough

For a straightforward view of recognised AI platforms and a general sense of trend direction, the native channel is sufficient.

Turning it on, annotating the launch date, and captioning the number as a floor covers most standard reporting needs.

When a Custom Channel Is Still Valuable

Reporting to a board, an executive team, or a client that expects a defensible figure calls for more rigour.

A custom channel group captures Perplexity and any other platform outside Google’s recognised list, and it remains the only way to apply AI classification to historical data.

How to Track AI Platforms Outside the Native Definition

A custom channel group with source rules for Perplexity, placed above the Referral rule, closes most of the coverage shortfall the native channel leaves behind.

GA4 evaluates channel rules in order, so placement affects the result even when the underlying regex pattern is correct.

Keep Your AI Traffic Numbers Honest

The GA4 AI Assistant channel is genuinely useful, but it has real limits.

It counts forward only from 13 May 2026, misses Perplexity and unrecognised platforms, and does not catch AI-referred sessions without a clean referrer.

AI Overviews and AI Mode still count as Organic Search, not AI Assistants.

Reading the channel’s number as a complete AI traffic figure leads to under-reporting the influence AI search already has on a business.

A native baseline, paired with supplementary tracking and Search Console data, gives a far more accurate account.

Accurate measurement is the foundation of good decision-making.

Tell No Lies provides the data engineering expertise and technical audits needed to build reporting that reflects what is actually happening across AI-driven traffic.

Contact us today to build a measurement setup you can trust.