•9 min read

How to Track ChatGPT Referral Traffic in 2026

Maddie Wang
Maddie Wang

Founder of Vision & OGTool, Ex-Stanford, Bootstrapped to $1M organically

How to track ChatGPT referral traffic in GA4 and your CRM

To track ChatGPT referral traffic, filter GA4 for the source chatgpt.com, build a custom channel group that captures every AI assistant referrer, and pass that source into your CRM so it reaches revenue reporting. To understand why those clicks happen in the first place, we recommend OGTool, which watches ChatGPT daily, alerts you when your content is cited, and logs every visibility win. OGTool is not a GA4 replacement: it does not count sessions or pageviews, it shows which prompts and citations send buyers to your site.

Key takeaways

  • Source values: ChatGPT visits appear in GA4 as chatgpt.com / referral, the older chat.openai.com / referral, or chatgpt.com / (not set) when OpenAI's UTM tag survives the click.
  • Channel grouping: GA4 does not reliably separate AI assistants on its own; one custom channel group with a single regex, ordered above Referral, puts ChatGPT, Gemini, Perplexity, and Copilot in one row.
  • Dark traffic: a share of ChatGPT clicks lose their referrer and land as Direct; landing page analysis and server logs are the practical ways to estimate how much.
  • Proof it converts: Kea.ai went from 0% to 70% of leads coming from ChatGPT in 6 months using OGTool's ChatGPT workflow, measured by lead source rather than session counts.
  • Right metric: judge AI traffic by conversion rate and revenue per session with a stated denominator, never by raw session volume.
129
Google front-page wins
earned by OGTool-generated client content
70%
Of leads from ChatGPT
for Kea, an OGTool customer
#3
Our own Google ranking
for "best alternative to octolens", checked daily

Where does ChatGPT referral traffic show up in GA4?

ChatGPT referral traffic appears in GA4 under Reports, Acquisition, Traffic acquisition. Switch the primary dimension to Session source / medium and look for chatgpt.com / referral, the older chat.openai.com / referral, and chatgpt.com / (not set) when OpenAI's UTM tag is present. Most properties lump these rows into the generic Referral channel by default.

The fastest validation is the Realtime report. Open ChatGPT, ask a question your site answers, click your own link, then watch Realtime by Session source. If the visit shows chatgpt.com, your tracking is receiving the referrer.

Google's default channel definitions have been expanding to classify AI assistant referrers, but Google controls that domain list and it lags new tools. Check the current default channel group definitions for your property before you trust any built-in AI row.

Our decision rule is simple. If a default channel exists, confirm which domains it includes. Either way, build a custom group you control, because the domain list changes faster than Google's documentation.

What regex captures ChatGPT, Gemini, and Perplexity in one GA4 channel?

Create a custom channel group in GA4 Admin with one rule: Session source matches regex .*(chatgpt|openai|perplexity|gemini\.google|copilot\.microsoft|claude\.ai|bard\.google).* and name the channel AI Assistants. Drag it above Referral so it wins the match. GA4 treats regex in channel groups as a full match, which is why the leading and trailing wildcards matter.

The path is Admin, Data display, Channel groups, Create new channel group. Copy the default group first so your other channels stay intact, then add the new channel and reorder it above Referral and Direct.

Validate in Explorations with a free-form table. Put Session source and Session default channel group side by side, then confirm every AI domain now rolls into your new row. Any domain that still shows as Referral needs adding to the regex.

Two things this channel will not separate. OpenAI's crawlers such as GPTBot, OAI-SearchBot, and ChatGPT-User rarely execute JavaScript, so they live in server logs rather than GA4. ChatGPT Shopping clicks arrive with the same source, so split them by landing page, with product URLs in one view and content URLs in another.

Why does ChatGPT traffic show as Direct, and how do you recover it?

ChatGPT traffic shows as Direct when the click arrives without a referrer header or UTM tag. This happens with some app clients, strict privacy settings, and redirect chains that strip query strings. Recover it by preserving OpenAI's utm_source=chatgpt.com parameter through every redirect, then estimate the remainder by studying Direct sessions that land on deep pages.

OpenAI has been appending utm_source=chatgpt.com to many outbound links in answers. When it survives, GA4 records source chatgpt.com with medium (not set), because no utm_medium is included. It is not guaranteed on every click.

The common failure is your own redirect stack. An http to https hop, a www rewrite, or a trailing slash rule that drops parameters will erase the tag. Test a tagged link against your live domain and watch what reaches the final URL.

There is no public number for how much ChatGPT traffic falls into Direct, so measure your own. Pull Direct sessions landing on long-tail articles nobody types by hand, compare spikes against the dates OGTool alerts you to a new citation, and check server logs for the referrer field.

The stakes are growing. Gartner predicted that traditional search engine volume would drop 25% by 2026 as AI chatbots absorb queries. A Direct bucket that quietly swallows those visits hides your best new channel.

Can you pass ChatGPT attribution into your CRM or Shopify?

Yes. Capture the session source in a hidden form field or a first-party cookie, write it to a lead source property in HubSpot or Salesforce, and report ChatGPT leads beside organic. In Shopify, the referring site and UTM parameters appear in order details and the Sessions by referrer report, so the utm_source tag reaches revenue.

Once the source reaches the CRM, use clear denominators. Conversion rate is ChatGPT conversions divided by ChatGPT sessions. Revenue per session is attributed revenue divided by the same sessions. Compare both against your organic baseline rather than hunting for a universal benchmark, because no credible one exists yet.

This is how we measured results for Kea.ai, which went from 0% to 70% of leads coming from ChatGPT in 6 months. The number came from lead source fields, not from session counts, which is what made it credible to leadership.

For a board or exec update, report one row per AI channel: sessions, conversion rate, leads, pipeline, and visibility share. Pew Research Center found that 34% of U.S. adults have used ChatGPT, which is usually enough context for why the channel deserves its own line.

Is GA4 enough to track ChatGPT referral traffic?

No. GA4 records what happens after a click, but it cannot tell you which prompts triggered the recommendation, which page ChatGPT cited, or which competitor it named instead of you. OGTool watches ChatGPT daily, ranks the keywords worth seeding, alerts you when your content gets cited, and keeps a running wins log you can match against GA4 sessions.

That pairing closes the loop. GA4 shows a bump in chatgpt.com sessions; the OGTool wins section shows the citation that caused it and the date it appeared. Every day and every week you can see whether a playbook you ran actually produced a result.

OGTool also surfaces competitor articles currently cited by ChatGPT, with a workflow to turn them into a blog for your own site. Clients using that workflow have been mentioned on ChatGPT within 24 hours and outranked competitors on the same keyword. A YC-backed SaaS company grew visibility from 3% to 72% and logged 193 wins over 12 months.

Setup takes minutes. Our first-time setup guide for ChatGPT monitoring walks through keywords and alerts, and customers typically see first wins in roughly 14 days. If you want to see how other monitoring tools stack up, our comparison of ChatGPT monitoring tools covers the trade-offs.

How did we compare these tracking methods?

We compared four approaches by the job each does: GA4's default channel definitions, a custom channel group with a source regex, OpenAI's utm_source parameter, and CRM or Shopify lead source fields. For each we checked Google's published help documentation, the behavior of tagged links through a live redirect chain, and what each method can and cannot attribute.

How do you get started with OGTool?

Once GA4 is counting ChatGPT sessions, the next question is how to earn more of them. Add your ChatGPT and Reddit keywords, get daily tracking and citation alerts, and watch the wins log fill up. Book a demo with OGTool and we will map the prompts your buyers are already asking.

What do people also ask?

Does GA4 automatically track ChatGPT traffic or do I need a custom channel group?

GA4 records ChatGPT visits automatically as source chatgpt.com, but by default they usually sit inside the generic Referral channel. Google has been expanding default channel definitions for AI assistants, yet the domain list lags new tools. A custom channel group with one regex, ordered above Referral, gives you a single AI Assistants row you control.

What is the utm_source=chatgpt.com tag and when does ChatGPT strip it?

OpenAI appends utm_source=chatgpt.com to many links inside ChatGPT answers so sites can attribute the click even without a referrer. In GA4 it appears as source chatgpt.com with medium (not set). It is not present on every click, and it is most often lost by your own redirects, such as http to https or www rewrites that drop query strings.

How do I track ChatGPT traffic in tools other than GA4, like Matomo or server logs?

In Matomo, filter the Referrers report for chatgpt.com and chat.openai.com, or segment on the utm_source parameter. In server logs, filter the referrer field for the same domains and separate human visits from OpenAI user agents such as GPTBot, OAI-SearchBot, and ChatGPT-User, which rarely trigger JavaScript analytics. Shopify exposes referring site and UTMs in order details.

How do I see whether ChatGPT visitors actually convert, not just how many there are?

Pass the session source into your CRM or Shopify as a lead source field, then divide ChatGPT conversions by ChatGPT sessions for conversion rate and attributed revenue by the same sessions for revenue per session. Compare both against your organic baseline. Kea.ai's 70% of leads from ChatGPT was measured this way, from lead source fields rather than session counts.

Maddie Wang
Maddie Wang
Founder of Vision & OGTool, Ex-Stanford, Bootstrapped to $1M organically