This is a practical setup guide for SaaS teams to isolate, measure, and act on the clicks ChatGPT sends to their website.

Updated by
Updated on Jul 03, 2026
ChatGPT clicks don't show up as their own channel in standard analytics because Google Analytics 4 does not, by default, treat AI platforms as a distinct traffic source category. Sessions from chatgpt.com typically land in the generic Referral channel, and sessions where the referrer header was stripped — common on mobile apps — land in Direct instead.
This matters for SaaS teams specifically because B2B buyers researching software are more likely than average users to be using ChatGPT on a phone between meetings, which increases the share of AI-influenced visits that lose their referrer entirely. Without deliberate configuration, a SaaS team can be receiving a meaningful and growing volume of ChatGPT-driven traffic while their dashboards show none of it.
Original insight: A useful gut check is to look at your Direct traffic trend line alongside your AI mention rate over the same period. If Direct traffic is climbing at the same time your brand's ChatGPT mention rate is rising, some portion of that "Direct" growth is very likely unattributed AI referral traffic hiding in plain sight.
The most reliable way to track ChatGPT clicks in GA4 is to create a custom channel group with a regex condition on the session source, since this recategorizes the traffic across every standard report rather than requiring a one-off filtered view. This is a one-time setup that applies going forward, though it does not automatically apply to historical data collected before the group is created.
GA4 also rolled out a native "AI Assistant" default channel group in mid-2026, which automatically classifies some sessions from recognized AI platforms without any manual setup. However, this native channel is not retroactive and depends on Google's own list of recognized AI referrers, which may lag behind newer or smaller platforms. Keeping a custom channel group active alongside the native one — rather than replacing it — is the more complete approach, since the two can be combined with an OR condition so a session is captured whether it matches your regex or Google's native classification.
Once ChatGPT traffic is isolated into its own channel, the most actionable next step is building an exploration that shows landing page performance by AI source, since this tells you not just how much traffic is arriving but which specific pages are actually being cited. In GA4's Explore section, a free-form exploration using session source as a dimension alongside landing page and engagement metrics will show which pages ChatGPT is sending users to — information that maps directly back to which content is currently earning citations.
This view is useful for two distinct purposes. First, it validates which pages are already working, so a SaaS team can identify what those pages have in common structurally and apply the same pattern elsewhere. Second, it surfaces landing pages that receive AI-referred traffic but perform poorly on engagement or conversion, which points to a landing-page problem rather than a visibility problem — the brand is being cited, but the page isn't converting the resulting visit.
Practical example: A SaaS team might find that a comparison page is receiving a growing share of ChatGPT-referred sessions but converting at half the rate of their pricing page. That gap suggests the comparison page needs a clearer next step for the visitor, not more content — a landing-page fix, not a visibility fix.
Even a correctly configured GA4 setup will undercount ChatGPT-driven traffic, because a meaningful share of AI-influenced visits never generate a trackable click at all. Two structural gaps are worth understanding so SaaS teams don't over-interpret what their dashboards show.
The first gap is referrer loss. Mobile app environments frequently strip referrer information when a user taps a link, which means a real ChatGPT-driven visit can land in Direct traffic with no way to trace it back to the originating platform through standard analytics alone. Since ChatGPT began appending UTM parameters to some citation links in 2025, a portion of this gap has narrowed, but it has not closed — plenty of sessions still arrive without any attribution signal.
The second, larger gap is zero-click influence. Analysis of AI search behavior has found that a large majority of AI Mode sessions end without any click to an external site at all, meaning a user can read a ChatGPT answer that names your brand, form an opinion, and never generate a session your analytics can see. That user may search your brand name directly later, and if they do, standard analytics will attribute the resulting session to branded organic or direct traffic — not to the AI answer that actually drove the decision.
Original insight: Because of this second gap, GA4 click tracking should be treated as a partial view of AI influence, not a complete one. A SaaS team that only tracks clicks is measuring the smaller, more visible slice of ChatGPT's effect on its funnel — mention-rate tracking is the layer that captures the larger, invisible slice.

Dageno AI complements GA4 click tracking by measuring the layer that happens before a click is even possible — whether ChatGPT mentions and cites your brand in the first place. Dageno AI provides the workflow from data monitoring → strategy → content generation → result attribution, connecting the mention-level data GA4 cannot see to the click-level data it can.
Data monitoring: Dageno AI runs real prompts against ChatGPT and other AI platforms on an ongoing basis, tracking mention rate, citation rate, and answer position — the upstream signals that determine whether a click ever becomes possible, regardless of what your GA4 dashboard shows.
Strategy: By comparing prompts where competitors are cited against prompts where your brand is absent, the platform identifies exactly which content gaps are suppressing both mentions and, downstream, clicks.
Content generation: The same workflow supports building content targeted at the highest-value gap prompts, so future GA4 reporting has more linked mentions to actually capture as referral traffic.
Result attribution: Because Dageno AI re-runs the same prompts over time, teams can pair rising mention and citation rates in Dageno with rising AI-referral sessions in GA4, confirming the two data sources tell a consistent story.
Get your website's GEO report!
Get started now - get it for free!>For a closer look at what drives the mention and citation side of this pipeline, see Dageno's guide to tracking brand mentions on ChatGPT, and for teams building this into a repeatable software evaluation, the best software to track brand mentions in AI responses covers the broader tool landscape.
Partially — GA4 introduced a native "AI Assistant" channel in 2026 that automatically classifies some sessions from recognized AI platforms, but it only works for sessions with intact referrer data and does not apply retroactively to historical data. A custom channel group with a regex condition is still recommended alongside it for more complete and historically consistent tracking.
This happens when the referrer header is stripped before the click reaches your site, which occurs frequently with mobile app browsing and with some in-app link handling. This traffic cannot be reliably reclaimed through analytics configuration alone, which is why Direct traffic trends are worth watching alongside AI visibility metrics as an indirect signal.
A working pattern needs to match the domains of the AI platforms relevant to your audience, typically including chatgpt.com, perplexity.ai, claude.ai, gemini.google.com, and copilot.microsoft.com at minimum, with additional platforms added as they become relevant to your category. Review and update the pattern periodically, since new AI platforms and domain changes are common.
You can use the same metrics, such as conversion rate and engagement rate, but expect the underlying user behavior to differ meaningfully between the two channels rather than assuming parity. Several independent studies have found ChatGPT-referred sessions show different engagement patterns than organic search, so treat any comparison as directional rather than exact.
Not necessarily — GA4 click tracking only captures the portion of AI influence that results in a trackable click, and a large share of AI-driven brand exposure never produces a click at all. Low or zero AI-referral sessions in GA4 could reflect either a genuine visibility gap or simply the zero-click nature of many AI search sessions, which is why prompt-level mention tracking is a necessary complement to click tracking.
At minimum quarterly, since new AI platforms emerge regularly and existing ones periodically change how they pass referrer or UTM data. Treat this as ongoing maintenance similar to reviewing UTM conventions, not a one-time setup task.
OpenAI Help Center – ChatGPT Search
Analytics Mania – How to Track and Report AI Traffic in GA4
GA4 Optimizer – GA4 AI Assistant Channel: How to Track Chatbot Traffic

Updated by
Richard
Richard is a technical SEO and AI specialist with a strong foundation in computer science and data analytics. Over the past 3 years, he has worked on GEO, AI-driven search strategies, and LLM applications, developing proprietary GEO methods that turn complex data and generative AI signals into actionable insights. His work has helped brands significantly improve digital visibility and performance across AI-powered search and discovery platforms.

Tim • May 21, 2026

Ye Faye • May 21, 2026

Richard • Jun 08, 2026

Tim • Jun 08, 2026