How to Track ChatGPT Referrals in GA4 (Google Analytics 4)

TL;DR

• AI‑driven referrals from ChatGPT are still tiny (around 0.19 % of traffic) but growing fast.

• Use GA4’s Traffic Acquisition report and custom Explorations or channel groups to isolate chatgpt.com / referral and other AI sources; add a regex filter to include Gemini, Perplexity, Copilot, Claude and similar domains.

• Free ChatGPT users don’t send referrer data, so some AI clicks appear as “Direct.” Add your own UTM parameters when sharing links in ChatGPT to ensure attribution.

• Track which landing pages AI tools cite, update your regex patterns as new AI platforms emerge, and keep creating authoritative, problem‑solving content---AI referrals may be small now, but they’re poised to become a meaningful channel.

---

AI-powered search is changing how we track website traffic. This guide shares top tips and answers common questions, so you can confidently track and understand referrals from ChatGPT.

1) Know how ChatGPT referral traffic works

Sometimes, ChatGPT includes links to websites in its answers. When someone clicks one of those links, OpenAI often adds a tag like utm_source=chatgpt.com to the URL. GA4 will then log that visit as Referral traffic, with the session source showing as chatgpt.com / referral.

But this doesn’t happen for every user. People using the free version of ChatGPT don’t send referrer data, so their visits show up as Direct traffic instead. Other AI tools---like Gemini, Perplexity, Copilot, and Claude---work in a similar way. Some of them even remove all referrer info.

Because of this, the number of AI referrals shown in GA4 is actually lower than the true number of visitors coming from AI tools.

2) How to check for ChatGPT referrals in GA4 (Google Analytics 4)

1. Open Reports → Traffic acquisition.

2. Above the table, change the dimension dropdown to Session source / medium.

3. In the search box, type chatgpt (or gpt). You’ll see rows like “chatgpt.com / referral”, indicating visits from ChatGPT. If nothing appears, you have no recorded referrals in the selected date range.

4. To see which pages were clicked, click the + next to the primary dimension and add Landing page + query string. GA4 will display a table showing the destination pages and the number of sessions from ChatGPT.

This quick check gives you a snapshot of AI traffic, but it’s limited. You can’t easily compare multiple AI sources or see trends over time. That’s where Explorations and channel groups come in.

3) Build a GA4 Exploration for AI referrals

GA4’s Explore section allows you to build custom reports. Here’s a repeatable approach:

1. Navigate to Explore and create a Blank exploration. Give it a name like “AI Traffic” and choose a long date range (e.g., “Last 12 months”).

2. Add the Session source (or Session source / medium) dimension and Landing page + query string. Add metrics such as Sessions, Engagement rate, Active Users, and Key Events.

3. Scroll down to Filters and set a condition on Session source. Choose Matches regex and paste a pattern that matches multiple AI domains, for example:

^https://(chatgpt\.com|gemini\.google\.com|perplexity\.ai|copilot\.microsoft\.com|claude\.ai|openai\.com)(/.*)?$

This pattern will capture sessions coming from ChatGPT, Gemini, Perplexity, Copilot, Claude and OpenAI. Adjust the regex to include other sources you see in your data or remove ones you don’t. GA4’s regex matching is case‑sensitive, so double‑check the source values in your property.

4. Add Landing page + query string to Rows, Session source to Columns and Sessions (plus any other metrics) to Values. You now have a table showing pages and sessions broken down by AI source. Switch the visualisation to a line chart to see trends over time.

Explorations are private to your user account but can be shared. They’re ideal for ad‑hoc analysis and pattern discovery.

4) Create an AI channel group for ongoing reporting

If you want AI referrals to appear alongside other marketing channels (Organic, Social, Paid Search, etc.), create a custom channel group:

1. Go to Admin → Data display → Channel groups. Click Create new channel group.

2. Name the group something like “AI Traffic Channels”. Click Add new channel, name it “AI Traffic” and set the Source condition to Matches regex. Use a pattern such as ^.*(chatgpt\.com|gemini\.google\.com|openai\.com|perplexity\.ai|copilot\.microsoft\.com).*

3. Save the new channel and drag it above Referral in the channel list. The order matters — GA4 assigns traffic to the first matching channel, so placing AI above Referral ensures AI visits aren’t lumped in with generic referrals.

4. After saving, open your Traffic acquisition report and select your new channel group from the dropdown. AI traffic will now appear as its own channel.

Custom channel groups apply retroactively to historical data and are visible to anyone with access to the property. They provide a consistent way to monitor AI traffic without manually filtering.

5) Use UTM parameters when sharing links in ChatGPT

ChatGPT often appends its own utm_source=chatgpt.com parameter, but you can override or supplement this by adding your own tracking parameters. If you share a link in ChatGPT or any AI tool, include descriptive UTM parameters such as:

https://www.example.com?utm_source=chatgpt&utm_medium=organic&utm_campaign=ai_referral

This ensures that GA4 records the session source precisely, even when referrer data is missing. Criterion B notes that using UTM parameters is one of the most effective ways to capture AI‑driven traffic.

6) Interpret and optimise your AI traffic data

1. Look beyond volume: AI referrals are still small but tend to come from users with high intent. Compare engagement rate, average engagement time and conversion metrics between AI traffic and other channels.

2. Identify landing pages: Use your exploration or channel group to see which pages AI tools link to. Optimise those pages with clear summaries, strong intros and structured headings so that AI models continue to cite them.

3. Monitor trends: Both the number of AI visitors and the list of AI domains will change quickly. Keep your regex patterns up to date and watch for new referral domains (e.g., gemini.google.com, claude.ai).

4. Consider brand mentions: AI conversations generate far more brand mentions than clicks. While those mentions aren’t yet measurable in GA4, they influence awareness and could lead to later visits via other channels. Use an AI visibility tool like Rankshift to discover how AI search platforms mention your brand.

7) Prepare for the future of AI-driven search

AI search tools and user behavior are still changing. Free users might never send referrer data, and some platforms may not give any referral info at all. To stay ready:

• Create helpful, expert content: Content that clearly solves problems is more likely to be linked by AI tools. Use natural language and long, specific keywords. AI search is more like a conversation.

• Let AI bots crawl your site: If you block AI bots, tools like ChatGPT or Gemini might not link to your pages. Check your robots.txt file and meta tags to make sure AI bots are allowed.

• Watch for spikes in direct traffic: If you notice sudden increases in direct traffic, compare it with ChatGPT mentions or viral content from AI tools. Some of that traffic may really be from AI sources, not true “direct” visits.

• Keep your team informed: Make sure others understand that some AI referrals will look like direct traffic and that the data will keep changing as AI platforms evolve. Keep updating how you track this traffic.

Key Takeaways

Tracking ChatGPT referrals in GA4 requires a multi-faceted approach. While AI-driven traffic is still relatively small, it’s growing rapidly and represents high-intent users. By implementing proper tracking methods, creating custom channel groups, and using UTM parameters, you can gain valuable insights into how AI platforms are driving traffic to your site.

Remember that the landscape is constantly evolving, so stay flexible with your tracking methods and keep an eye on emerging AI platforms that might become significant traffic sources in the future.