How to Track AI Traffic in Google Analytics and Search Console

Why AI Traffic Tracking Matters

If you cannot measure it, you cannot improve it. That applies to AI traffic more than almost any other channel right now, because AI traffic is growing rapidly but remains largely invisible in default analytics setups.

Semrush's 2026 AI traffic report found 357% year-over-year growth in AI-referred sessions across their tracked site panel. That is enormous growth. But most of those site owners had no idea it was happening because it was buried in their direct traffic or miscategorized as generic referral traffic.

Beyond just knowing the number, tracking AI traffic lets you answer the questions that matter: Which pages are getting cited by AI tools? Which queries are driving AI-referred visits? What is the conversion rate of AI-referred traffic compared to organic or paid? Is your AI SEO work actually moving the needle?

Without proper tracking, you are flying blind on a fast-moving channel. This guide sets up the tracking infrastructure you need.

The Core Problem With AI Traffic Attribution

The fundamental problem is that different AI tools handle referrer data differently, and several of them send no referrer information at all.

Here is how the major AI tools handle referrers:

AI Tool	Referrer Behavior	Where It Appears in GA4
Perplexity	Passes perplexity.ai as referrer	Referral: perplexity.ai
ChatGPT web	Often strips referrer	Direct (no source)
Bing Copilot	Passes bing.com as referrer	Referral: bing.com
Google AI Overviews	Passes google.com	Organic: google / organic
You.com	Passes you.com as referrer	Referral: you.com
Claude.ai	Inconsistent, often strips referrer	Direct or referral: claude.ai
Phind	Passes phind.com as referrer	Referral: phind.com
Gemini	Often passes google.com	Mixed with organic Google

The worst offender is ChatGPT, which is also the highest-volume AI tool. When ChatGPT cites your site and a user clicks through, GA4 often records that session as direct. This means your direct traffic channel is currently inflated with unattributed AI referrals.

There is no perfect solution to this problem. But there are good approaches that capture 70 to 80% of AI-referred traffic accurately. Here is how to set them up.

Segmenting AI Referral Traffic in GA4

The first step is to identify what AI-referred traffic you can already see in your existing GA4 data. Here is the process:

Step 1: In GA4, go to Reports, then Acquisition, then Traffic Acquisition.

Step 2: Set the primary dimension to "Session source/medium."

Step 3: Use the search bar above the table to filter for known AI domains. Search for each of these in turn:

• perplexity.ai / referral
• you.com / referral
• phind.com / referral
• claude.ai / referral
• chat.openai.com / referral
• copilot.microsoft.com / referral
• bing.com / referral (note: includes non-AI Bing referrals)

Step 4: Note the session counts and user behavior for each source (engagement rate, pages per session, conversion rate). Compare these to your overall site averages. AI-referred traffic often shows higher engagement because users who click through from an AI citation are already pre-qualified by the AI's answer.

Step 5: Save the traffic acquisition report as a saved card or export it to a spreadsheet. Set a calendar reminder to repeat this review monthly.

This gives you visibility into the AI-referred traffic that is already showing up in your data. The next step is to create a channel group that automatically segments these sources.

Creating a Custom AI Channel Group in GA4

GA4's channel groups let you define custom traffic categories that apply automatically to all incoming sessions. Creating an "AI Traffic" channel group is a one-time setup that makes AI referrals permanently visible in your standard acquisition reports.

Here is how to set it up:

Step 1: In GA4, go to Admin (the gear icon at the bottom left), then Data Display, then Channel Groups.

Step 2: Click "Create new channel group."

Step 3: Name the channel group "AI Traffic" or "AI Referrals."

Step 4: Click "Add new channel" and create channels for each major AI source. Use "Session source" as the dimension with "contains" as the condition. Add the following sources as separate channels or combine them into one:

• perplexity.ai
• you.com
• phind.com
• claude.ai
• chat.openai.com
• copilot.microsoft.com
• gemini.google.com
• poe.com

Step 5: Save the channel group. It will now appear as an option in your channel dimension dropdown across all GA4 reports.

Once your channel group is active, go to Reports, Acquisition, Traffic Acquisition, and switch the channel grouping to your new "AI Traffic" group. You will immediately see a cleaner breakdown of AI-sourced sessions.

GSC AI Overviews Filter Walkthrough

Google Search Console added a dedicated AI Overviews filter in late 2024. This is the most direct way to see your Google AI Overview performance, showing you exactly which queries trigger AI Overviews where your site is cited.

The walkthrough:

Step 1: Open Google Search Console and select your property.

Step 2: Click "Search results" in the left navigation under Performance.

Step 3: At the top of the report, click the "Search type" filter button (it defaults to "Web").

Step 4: In the filter dropdown, select "AI Overviews." The report will refresh to show only queries and pages associated with AI Overview citations.

Step 5: Review the Queries and Pages tabs. Queries shows you which user searches trigger AI Overviews where you are cited. Pages shows you which of your pages are being cited.

Key metrics to track in the GSC AI Overviews report:

Impressions: How many times your site appeared as a citation source in AI Overviews. This is your AI Overview "reach."

Clicks: How many times users clicked from an AI Overview citation to your site. This is your direct AI Overview traffic from Google.

CTR: Click-through rate from AI Overview impressions. This tends to be lower than traditional organic CTR because many users get their answer from the AI Overview and do not click through. A CTR of 2 to 5% is common for AI Overview citations.

Position: Your relative position among the source cards shown below the AI Overview. Position 1 is the first cited source. Higher positions (lower numbers) generally correlate with higher CTR.

Export this data monthly and track it over time. After you make content or schema improvements, watch for impression growth in the 2 to 4 weeks following the change.

Bing Webmaster Tools for Copilot Insights

Bing Webmaster Tools (BWT) provides insights into how your site performs in Bing Search, which is directly used by Microsoft Copilot. If Copilot is citing your content, it is usually because your content ranks well in Bing.

How to use BWT for AI traffic insights:

Set up BWT if you have not already. Go to bing.com/webmasters and verify your site. The process is similar to GSC verification and takes about 10 minutes.

Check the Search Performance report. Look at which queries are bringing Bing traffic to your site. These are the same queries for which Copilot might be citing you, since Copilot draws from Bing's search index.

Check the Index Freshness report. This shows how recently Bing crawled your pages. Pages that were crawled recently are more likely to be cited in Copilot responses for time-sensitive queries.

Look for "Copilot" in your GA4 referrals. Copilot-referred traffic appears in GA4 with the source copilot.microsoft.com. Cross-reference the queries BWT shows with the Copilot referral traffic in GA4 to build a picture of which Bing-ranking queries are driving Copilot citations.

Tracking Brand Mentions Manually (Weekly ChatGPT and Perplexity Checks)

Manual brand mention checks are low-tech but give you ground truth about whether AI models are actually citing your brand. No tool can fully replace direct testing.

Here is the weekly routine to run:

The ChatGPT brand check (5 minutes). Open ChatGPT (with web browsing enabled, or using a current model that has recent knowledge). Ask: "What is [your brand name]?" and "What does [your brand name] do?" Note how accurately ChatGPT describes you, whether it cites your website, and what it gets right or wrong. Wrong information in ChatGPT is an E-E-A-T and entity clarity problem. Fix it by improving your About page, Organization schema, and Wikipedia presence.

The Perplexity topic check (10 minutes). Open Perplexity and search your top 5 to 10 target queries. See which sources Perplexity cites for each. Is your site in the citations? If not, which sites are? Click through to those competitor pages and analyze what they have that your pages do not. This is a direct competitive intelligence exercise you can do for free.

The Gemini check (5 minutes). Open Gemini and ask your top 3 to 5 queries. Gemini is powered by Google's AI and draws from Google's index, so strong Google rankings usually translate to Gemini citations. If you rank well on Google for a query but are not being cited by Gemini, content structure and quotability are likely the issue.

Record the results of each weekly check in a simple spreadsheet. Track whether citations are increasing or decreasing over time. This manual tracking takes about 20 minutes per week and gives you data no tool can replicate: direct observation of what AI systems actually say about you and your topics.

Setting Up UTM Parameters for AI Tools

UTM parameters are URL tags that pass custom attribution data to Google Analytics. They are the most reliable way to track clicks from specific AI sources, because they bypass the referrer-stripping problem entirely.

UTM parameters look like this appended to a URL:

https://yoursite.com/blog/your-article
  ?utm_source=perplexity
  &utm_medium=ai-referral
  &utm_campaign=content-quotability-guide

When a user clicks that URL, GA4 records the session with exactly the source, medium, and campaign you specified, regardless of what the referring page's server sends in the referrer header.

Where to use UTM-tagged URLs for AI traffic:

Your llms.txt file. In your llms.txt file, include UTM-tagged URLs for your key pages. When AI crawlers read your llms.txt and include your pages in their training or retrieval, any resulting clicks will carry your UTM parameters. Add utm_source=llms-txt and utm_medium=ai-referral.

Schema markup sameAs and url fields. In your Organization and Article schema, the url field can include a UTM-tagged URL. AI systems that read schema markup may use these URLs as citation links.

AI-facing content pages. If you create content specifically optimized for AI citation (landing pages, FAQ pages, resource pages), use UTM-tagged canonical links to track whether those pages drive AI referrals.

To build UTM-tagged URLs quickly, use Google's Campaign URL Builder at ga-dev-tools.google.com/campaign-url-builder. Always use lowercase, no spaces, and consistent naming conventions across your UTM parameters. "perplexity" not "Perplexity." "ai-referral" not "AI Referral." Inconsistency in UTM naming creates fragmented data.

Measuring AI Traffic ROI

Traffic without revenue impact is just a vanity metric. Here is how to connect AI traffic to actual business outcomes.

Set up conversion tracking in GA4 first. Before you can measure AI traffic ROI, you need conversions defined in GA4. These might be form submissions, free trial signups, email newsletter signups, product purchases, or phone calls. If you do not have conversion events set up, do that before anything else.

Compare AI traffic conversion rates to other channels. In GA4, go to Advertising, then Attribution, then Conversion paths. Filter by your AI Traffic channel group (once you have it set up). Compare the conversion rate of AI-sourced sessions to organic search, direct, and paid traffic. AI traffic often converts at above-average rates because it is highly pre-qualified: the AI model already answered the user's question, and the user clicked through wanting more or ready to act.

Track assisted conversions. AI citations often contribute to conversions without being the last touch. A user might first find you via Perplexity, then return via direct or branded search, then convert. GA4's attribution reporting shows these multi-touch paths. Look for AI Traffic in the assisted conversions column to see its full impact.

Calculate revenue per AI citation. If you have ecommerce tracking or goal values set up, GA4 will show you revenue by channel. Divide AI-channel revenue by the number of AI-referred sessions to get revenue per AI-referred session. Compare this to your organic and paid benchmarks. This gives you a dollar value for each AI citation.

Track branded search increases as a proxy. AI citations increase brand awareness. One imperfect but useful signal is branded search volume: if Perplexity and ChatGPT are citing your brand, more users will search for you by name. Watch your branded keyword impressions in GSC. A growing branded search trend correlates with growing AI citation volume, even if you cannot attribute it directly.

What Good AI Traffic Metrics Look Like in 2026

What should you actually be aiming for? Here are the benchmarks based on data from sites that have invested in AI SEO optimization across different niches in 2026.

AI referral traffic as a percentage of total sessions. Early-stage AI SEO investment: 1 to 3%. Active optimization with schema, quotability, and content structure improvements: 5 to 10%. Sites that have built strong topical authority and AI-optimized content libraries: 10 to 25% in competitive informational niches.

GSC AI Overview impressions per month. Depends heavily on your content volume and niche. A site with 20 to 50 optimized blog posts should be seeing 5,000 to 50,000 AI Overview impressions per month. If you are well below this with a content library of that size, schema and content structure are likely the bottleneck. Run the AI SEO Audit to check.

AI Overview CTR. Average CTR from AI Overview impressions is 2 to 5%. If yours is consistently above 5%, your source card descriptions are compelling and your brand is recognized. Below 2% suggests your source card text needs improvement (it is often pulled from your meta description).

Engagement rate for AI-referred sessions. AI-referred sessions typically show engagement rates of 60 to 75% in GA4, compared to industry averages of 50 to 60% for organic search. Higher engagement is expected because AI-referred users are pre-screened by the AI's answer. If your AI traffic engagement rate is below your organic search rate, the page AI is sending people to may not match what the AI cited you for.

Week-over-week AI citation growth after optimization. After implementing schema, quotability improvements, and content updates, expect to see measurable AI Overview impression growth within 2 to 4 weeks. Perplexity citation changes often appear within 1 to 2 weeks. ChatGPT changes tied to training data take longer, sometimes months.

Start your AI traffic measurement journey by running the free AI SEO audit on your site. It gives you a baseline score across schema, content, and technical factors so you have a starting point to measure improvements against.

Frequently Asked Questions About AI Traffic Analytics