Gemini UTM Parameters: Track Gemini Traffic in GA4 (2026)

Q: How do I see Gemini traffic in GA4?

Start by checking GA4's AI Assistant channel for visits attributed to Gemini. Then review session-source data for values containing 'gemini' to identify tagged visits that may appear elsewhere. A custom channel group that combines relevant referrer and source rules can provide a more complete view of Gemini-attributed sessions.

Q: Should I add utm_source=gemini to my own URLs?

No. Publishers should not add Gemini-specific UTM parameters to their own canonical URLs simply to identify AI traffic. Doing so can incorrectly attribute visits from other channels and interfere with normal referrer-based attribution. Keep canonical URLs clean and use the referral or campaign information supplied during the actual visit.

Q: Are Google AI Overviews and AI Mode clicks tagged too?

There was no confirmed equivalent UTM-tagging behavior for Google AI Overviews or AI Mode clicks in the information available in early October 2026. Teams should avoid assuming that Gemini's observed tagging behavior applies to other Google AI search surfaces and should validate attribution using GA4, Search Console, and server-side data where available.

Q: Why did my Direct traffic drop after September 30?

One possible explanation is that some Gemini visits previously classified as Direct are now receiving identifiable referral or campaign attribution. Compare Direct and Gemini-attributed sessions by landing page before and after September 30, 2026, and use server logs or other first-party evidence where available before concluding that Gemini tagging caused the change.

Written byApoorv SharmaCo-founder, DerivateX

Reviewed byPawan BhargavSr. Content Writer, DerivateX

Published Oct 7, 2026Updated Oct 6, 2026

9 min read

Gemini UTM parameters started showing up on outbound Gemini links at the end of September 2026, and that gives B2B SaaS teams a cleaner way to count visits from Google’s assistant. Until now, a large share of those visits could fall into Direct in Google Analytics 4 (GA4), which makes Gemini look smaller than it is in budget conversations. The change is real but undocumented, so the right move is to verify it in your own data before you report on it. This guide covers what changed, what is still unknown, and the setup to finish this week.

TL;DR

A Reddit user spotted UTM tags on Gemini links around September 30, 2026, and Google’s John Mueller replied that he sees them too, as reported by Search Engine Journal.
Google has not documented when Gemini adds the tags, which answer types carry them, or the full parameter set, so treat early numbers as partial coverage.
OpenAI’s Publishers and Developers FAQ already states that ChatGPT adds utm_source=chatgpt.com to referral URLs from ChatGPT search, which means both major assistants can now be tracked without relying on the referrer alone.
GA4’s AI Assistant default channel classifies sessions by referrer and excludes Google AI Overviews and AI Mode, so clicks from those two surfaces still count as organic search.
A reported “AI Overview tracking parameter” from October 2, 2026 turned out to be a Chrome extension rewriting links, according to Search Engine Roundtable, so verify any new parameter in a clean browser before you build reports on it.
The practical work this week is a baseline export, a custom AI channel group, UTM capture in your CRM, and a server log check for tagging coverage.

What Changed With Gemini Links?

Gemini began appending UTM parameters to some outbound links around September 30, 2026. A Reddit user flagged it, Google’s John Mueller confirmed he could see the tags, and Google has published no documentation. Tagged links let analytics tools identify Gemini visits even when the browser or app strips the referrer.

Roger Montti’s report for Search Engine Journal quotes Mueller’s reply in full: “Nice. I see it too. Doesn’t Gemini have a referrer too though? I thought you could already track them.” Mueller later added that he would forward reproducible referrer examples to the team and that it “would be great to have these retained (in addition to the UTM-tagging).” That second quote matters because it suggests Google sees UTM tags and referrers as complementary signals.

In the same thread, a Reddit user claimed that referrer pass through from the Gemini mobile app runs around 9 percent and that Android assistant invocations strip it entirely. That figure comes from a community comment, and neither Google nor SEJ verified it. Treat it as a hypothesis to test in your own logs. Do not quote it to leadership as a measured rate.

Why Does This Matter for B2B SaaS Pipeline?

Gemini visits that arrive without a referrer land in Direct, so pipeline influenced by Gemini gets credited to brand, bookmarks, or nothing at all. UTM tags give marketing teams a second way to count those sessions, which makes it possible to compare Gemini against ChatGPT and organic search with real session and conversion data.

The budget consequence is simple. If one assistant is tagged and another is mostly invisible, the visible one wins the investment argument by default. That bias already shaped how many teams read AI referral share reports, where ChatGPT dominates partly because its traffic has been easier to attribute.

For a B2B SaaS team, the cost of misattribution shows up at the deal level. A buyer who asked Gemini for a shortlist, clicked your comparison page on a phone, and booked a demo two days later may show up as Direct first touch and branded search last touch. Better tagging will not solve journeys that span several sessions on its own, but it gives you a cleaner first touch to work with.

If Gemini is a priority channel for your category, the Gemini SEO work and the measurement setup should move together, because visibility you cannot measure rarely survives a budget review.

What Is Still Unknown About Gemini UTM Tags?

Google has not said which Gemini surfaces add tags, whether answers built from training data carry them, or whether the parameter values will stay stable. ChatGPT tags links from web search answers, and Gemini may follow a similar rule, but nobody outside Google has confirmed that. Expect partial coverage until Google documents the behavior.

Three open questions decide how much you can trust the new numbers:

Trigger conditions. SEJ notes that ChatGPT only tags links tied to answers grounded in web sources. Gemini’s rules are unknown, so some Gemini visits will still arrive untagged.
Channel placement in GA4. GA4 normally gives manual campaign parameters priority over referrer detection. If Gemini’s tags set a medium other than ai-assistant, tagged sessions could move out of the AI Assistant channel and into Referral or Unassigned, so check where they land before you trust the channel report.
Stability. Undocumented behavior can change without notice. Keep the raw source and medium values in your exports so you can reclassify history if Google changes the format.

How Should You Set Up Gemini Tracking in GA4 This Week?

Start with a baseline export, then build a custom channel group that combines GA4’s AI Assistant channel with source rules for Gemini and other assistants. Capture the first touch source in your CRM and compare GA4 against server logs. The setup takes a few hours and gives you a defensible AI line in reporting.

1. Export a before and after baseline

In GA4 Explorations, pull sessions and key events by session source, session medium, and landing page for September 1 to 29 and for September 30 onward. Filter for any source containing “gemini” and separately for Direct sessions landing on deep pages such as comparison, pricing, or integration URLs. A drop in deep page Direct sessions alongside a rise in Gemini sessions is the pattern you are looking for.

2. Build a custom AI channel group

Default channel groups cannot be edited in GA4, so create a custom channel group under Admin, then Data display, then Channel groups. Add an AI Search channel whose rule matches Default channel group exactly equal to AI Assistant, or session source matching a regex for the assistants you care about, such as gemini, chatgpt, perplexity, claude, and copilot. Place it above Referral so tagged assistant visits are not swallowed by the referral rule.

3. Capture the first touch in your CRM

Store utm_source, utm_medium, and the landing page in a first party cookie on the first visit, then write them into hidden form fields when the visitor converts. This is the step that turns an analytics curiosity into pipeline attribution. The HubSpot prompt to pipeline attribution guide covers the property set and deal stamping in detail.

4. Check tagging coverage in server logs

Consent banners and ad blockers drop a share of GA4 hits, so your server or CDN logs are the better coverage check. Count requests carrying a Gemini UTM, then count requests with a Gemini referrer and no UTM. The ratio between the two tells you how complete Gemini’s tagging is for your site.

5. Keep tagged URLs out of the index

Never add Gemini UTMs to your own links, internal navigation, or sitemaps, because that labels every visitor from every source as Gemini traffic. Make sure every page carries a self referencing canonical to the clean URL so parameter versions do not compete in search. This is basic hygiene, and it matters more now that assistants are appending parameters at scale.

Where Do AI Overviews and AI Mode Fit?

Clicks from Google AI Overviews and AI Mode are excluded from GA4’s AI Assistant channel and count as organic search. Search Console’s Generative AI performance report shows impressions for both surfaces combined. There is still no confirmed click level tag for either surface, and the October 2 parameter sighting came from a browser extension.

The Search Engine Roundtable update is a useful warning. SEO consultant Tom Pool first saw a st_source=ai_overview parameter on AI Overview links, then traced it to a Chrome extension rewriting URLs, likely to collect data. Any new parameter that appears in your reports should be reproduced in a clean browser profile before anyone builds a dashboard on it.

For Google’s own AI surfaces, the measurement stack stays the same for now: Search Console impressions by page, organic landing page trends, and prompt level visibility tracking. The Search Console generative AI report guide explains what that report can and cannot tell you.

How Should You Report AI Assistant Traffic to Leadership?

Report three lines each month: tagged assistant sessions by source, untagged assistant sessions identified by referrer, and pipeline with an assistant as first touch. Keep Google AI Overviews and AI Mode in a separate organic line. Label Gemini numbers from before September 30 as undercounted so trend lines are read correctly.

Leadership does not need the parameter mechanics. They need to know which assistants send buyers, whether those buyers convert, and how confident the numbers are. A short confidence note on each line, such as “partial tagging, undocumented by Google,” keeps the report honest and protects the channel from being cut on bad data.

If you want a fuller model for connecting AI visibility to revenue, the AI search ROI framework lays out the metrics DerivateX uses with clients. Teams that want a starting point can run a free AI visibility audit to see where Gemini, ChatGPT, and Google already mention them.

How DerivateX Approaches AI Referral Measurement

DerivateX treats assistant traffic as a set of separate channels, each with its own visibility, click, and pipeline line. Gemini, ChatGPT, and Google’s AI surfaces get measured on their own terms because each one passes different signals. Undocumented changes like Gemini’s tags get verified in logs before they reach a client report.

That approach runs through every GEO engagement, where measurement setup happens in the first month so later visibility gains can be tied to pipeline. Engagement options and scope are listed on the pricing page, and published outcomes are in the case studies and on Clutch.

FAQ

Does Gemini add UTM parameters to links?

Yes, Gemini has been observed adding UTM parameters to some outbound links since about September 30, 2026. Google’s John Mueller said he sees them too, but Google has not documented when the tags appear or which values they use. Expect partial coverage and verify the behavior in your own analytics.

How do I see Gemini traffic in GA4?

Check the AI Assistant default channel first, because GA4 assigns Gemini visits there when it detects a matching referrer. Then filter session source for “gemini” to catch tagged visits that may land elsewhere. A custom channel group that combines both rules gives the most complete view of Gemini sessions.

Should I add utm_source=gemini to my own URLs?

No, you should never add Gemini UTMs to your own links or pages. Doing so labels visitors from every source as Gemini traffic and overrides GA4’s referrer detection. Gemini surfaces your canonical URL, so keep canonicals clean and let Gemini or the referrer identify the visit.

Are Google AI Overviews and AI Mode clicks tagged too?

No confirmed tagging exists for AI Overviews or AI Mode clicks as of early October 2026. GA4 counts those clicks as organic search, and Search Console reports their impressions together in the Generative AI performance report. An AI Overview parameter reported on October 2 came from a Chrome extension.

Why did my Direct traffic drop after September 30?

A drop in Direct sessions on deep pages after September 30 may mean Gemini visits are now being credited correctly through UTM tags. Compare Direct and Gemini sessions by landing page before and after that date. Confirm the shift in server logs before you attribute the change to Gemini.