How to Track AI Search Traffic in 2026: ChatGPT, Gemini, Copilot and Google AI Overviews

The Short Answer

To track AI search traffic, create an "AI assistants" channel in Google Analytics 4, make sure calls, forms and bookings are tracked as key events, check Bing Webmaster Tools for Copilot citations, and ask every new lead how they found you. ChatGPT tags the links it sends with utm_source=chatgpt.com, so those visits are easy to identify. Google's AI Overviews and AI Mode are counted inside your regular Search Console data rather than reported separately.

No tool shows all of your AI visibility. Many answers mention a business without sending a click, and you'll never see those conversations. The goal is not a perfect number. It's enough signal to know whether AI discovery is producing customers, and whether that's changing.

This guide is about measurement. For how to improve your visibility in the first place, see our companion guide, AI Search Optimization: How to Get Found by ChatGPT, Gemini & Bing.

Why AI Visibility Is Hard to Measure

Traditional search gave you a tidy loop: impressions, clicks, rankings, all in one report. AI answers break that loop in four ways:

That doesn't make measurement pointless. It means you combine several partial signals, and judge the combination.

The Five Layers of AI Discovery

Every AI referral passes through the same sequence. At each layer, ask two questions: what has to be true, and what can you see?

1. L1

   Access: can AI systems reach your pages?

   Visible

   AI search features rely on crawlers. OpenAI says sites opted out of OAI-SearchBot "will not be shown in ChatGPT search answers, though can still appear as navigational links" [5]. Perplexity recommends allowing PerplexityBot "to ensure your site appears in search results" [7]. For Google's AI features, Googlebot is the crawler that matters [1].

   What you can see: your robots.txt file, and your server or CDN logs, which show which crawlers visited and whether they were blocked.

2. L2

   Index: are your pages in the search indexes?

   Visible

   Google says that to appear as a supporting link in AI Overviews or AI Mode, "a page must be indexed and eligible to be shown in Google Search with a snippet," and that "there are no additional technical requirements" [1]. Microsoft reports Copilot's use of your pages through Bing Webmaster Tools, so Bing indexing matters too.

   What you can see: indexing status in Google Search Console and Bing Webmaster Tools.

3. L3

   Citation: do answers use and cite you?

   Partly visible

   This is the layer most businesses want to see and the one with the least data. Bing Webmaster Tools now offers an AI Performance report showing how often your pages are cited "across Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations" [8]. Google, OpenAI and Perplexity don't offer an equivalent citation report for site owners in the documentation we reviewed.

   What you can see: Bing's citation data, plus what you observe by asking assistants yourself.

4. L4

   Click: do people visit your site?

   Partly visible

   ChatGPT "automatically includes the UTM parameter utm_source=chatgpt.com in referral URLs" [4]. Other assistants can appear as referrals, and some visits arrive with no referrer at all. Clicks from Google's AI features are counted as Google Search traffic [1].

   What you can see: AI referral sessions in Google Analytics, once you group them into their own channel.

5. L5

   Lead: do those visitors become customers?

   Visible, if you ask

   This is the layer that matters to the business, and it's the easiest to measure if you set it up. Track calls, forms and bookings as key events, and record where every lead says they found you.

   What you can see: key events by channel in Google Analytics, and self-reported sources in your CRM.

What Each Platform Actually Shows You

Here's what the platforms document, as of September 2026. These details change often, so check the linked documentation before relying on them.

How to Set Up AI Traffic Tracking: 7 Steps

These steps take an afternoon for most small business websites, and most of them are free.

1. 01

   Check that AI crawlers can reach your site

   Open yoursite.com/robots.txt. Make sure it doesn't block Googlebot, Bing's crawler, OAI-SearchBot or PerplexityBot, unless that's a deliberate choice. Also check your hosting or CDN's bot protection, since firewalls can block crawlers that robots.txt allows. Google lists "ensuring that crawling is allowed in robots.txt, and by any CDN or hosting infrastructure" among its best practices [1].

   Decide separately about training crawlers. OpenAI's GPTBot collects content that "may be used in training" its models, and blocking it is a separate choice from blocking OAI-SearchBot [5].

2. 02

   Verify Search Console and Bing Webmaster Tools

   Both are free. Search Console shows clicks, impressions and positions for your pages in Google Search [2], including clicks from AI Overviews and AI Mode within the overall totals. Bing Webmaster Tools shows your Bing performance and, through AI Performance, how often Copilot and Bing's AI summaries cite your pages, which of your pages are cited and the "grounding queries" used to find them [10]. Submit your sitemap to both. If pages are missing from Google entirely, start with Why Isn't My Website Showing Up on Google?

3. 03

   Create an "AI assistants" channel in Google Analytics 4

   By default, visits from AI assistants are mixed in with other referrals. Google's own help documentation walks through creating a custom channel group with an "AI assistants" channel [3]:

   1. In Admin, open Channel groups and create a new custom channel group.
   2. Add a channel named AI assistants, with a condition on source that matches regex.
   3. Enter a pattern listing the AI sources you want to include. Test it with the tool below first.
   4. Move AI assistants above Referral in the channel list, so AI visits are classified before the referral rule catches them. Google notes you can also move it above other channels, like Organic Search [3].
   5. View it in Acquisition › Traffic acquisition, using your new channel group as the dimension.

   Custom channel groups "can be applied to your reports retroactively" [3], so you'll see past AI visits too.

4. 04

   Track the actions that make you money

   A visit from ChatGPT isn't worth much unless it becomes a call, form or booking. Mark those actions as key events in Google Analytics [11], then compare key events by channel. With small numbers, be careful drawing conclusions. Our conversion rate guide includes a calculator that shows how much a rate based on a few visits can move by chance.

5. 05

   Ask every lead how they found you

   This is the most underrated AI measurement tool. Add "How did you hear about us?" to your forms and intake calls, with an option such as "ChatGPT or another AI assistant," and record the answer in your CRM. It captures what analytics can't: the person who got your name from an AI answer and then searched for you, called you or walked in.

6. 06

   Check what AI assistants say about you, monthly

   Write down five to ten questions your customers actually ask, such as "best family dentist in Henderson" or "how much does a roof inspection cost." Each month, ask them in ChatGPT, Gemini, Copilot and Perplexity, and record whether you're mentioned, whether you're cited with a link, and whether the details are correct.

   Treat this as a spot check, not a ranking report. Answers vary by person, location, wording and time, and OpenAI itself warns that "search results and citations can be incomplete, outdated, or incorrect" [9]. Inaccurate details are the most useful finding: they usually point to outdated information on your website or listings.

7. 07

   Review the signals together, not one at a time

   Once a month, put four numbers side by side: AI assistant sessions and key events in Google Analytics, Copilot citations in Bing Webmaster Tools, leads who named an AI assistant, and your spot-check results. Note any content changes you made. No single number tells you much; a consistent direction across several of them does.

Test Your AI Assistants Channel Pattern

Google Analytics regex is "full regex and case sensitive" by default [12]: the pattern has to match the whole source value. Google's example pattern for the AI assistants channel is written to be broad. It includes any source ending in "ai" and any source containing "gpt" or "gemini," which can pull in sites that aren't AI assistants at all. Test a pattern against real source names before you use it. This tester runs in your browser and stores nothing.

Myths and Limits of AI Search Measurement

From Measurement to Improvement

Measurement tells you where you stand; it doesn't change it. Google's advice for its AI features is the same as for Search: meet the technical requirements and focus on "helpful, reliable, people-first content" [1]. In practice, that means clear service pages, accurate business details everywhere you're listed, genuine reviews and direct answers to the questions customers ask. We cover those in AI Search Optimization and, for local businesses, How to Rank Higher on Google Maps. And AI discovery is one channel among many: see how it fits with the rest in How to Get More Leads.

The Bottom Line

You can't see everything AI assistants say about your business, but you can see much more than most businesses do. Make sure AI crawlers can reach you, give AI referrals their own channel, track the actions that produce revenue, check Bing's citation data and ask every lead where they found you. Then judge AI discovery the way you'd judge any channel: by the customers it produces.