How to Track ChatGPT, Perplexity, and AI Assistant Traffic to Your Website (2026)
AI assistants now send real visitors to websites. Learn how to identify ChatGPT, Perplexity, Gemini, Copilot, and Claude referral traffic, measure its quality, and see what those visitors do.
Search used to mean Google. In 2026, a growing share of discovery happens inside AI assistants—ChatGPT, Perplexity, Gemini, Copilot, Claude—which increasingly cite and link to websites. Those clicks are real visitors, often with high intent, but many analytics setups lump them into "referral" or even "direct."
Here's how to identify AI assistant traffic, measure its quality, and see what those visitors actually do.
Table of Contents
- Why AI Traffic Is Easy to Miss
- How to Identify Each Assistant
- Setting It Up in Your Analytics
- Measuring Quality, Not Just Volume
- What AI Visitors Do Differently
- Key Takeaways
Why AI Traffic Is Easy to Miss
- Small volume spread across many referrer domains.
- Missing referrers when links are opened from apps or copied.
- Default channel groupings that don't have an "AI" bucket.
- Mixed signals—some assistants add UTM parameters, others don't.
How to Identify Each Assistant
| Assistant | Typical signals |
|---|---|
| ChatGPT | Referrer chatgpt.com; links often carry utm_source=chatgpt.com |
| Perplexity | Referrer perplexity.ai |
| Google Gemini | Referrer gemini.google.com |
| Microsoft Copilot | Referrer copilot.microsoft.com |
| Claude | Referrer claude.ai |
Referrer behavior changes as products evolve, so review your raw referrer list monthly for new domains.
Some AI traffic will look like direct
Visitors who copy a URL from an assistant's answer, or open it from a desktop or mobile app that strips referrers, arrive with no referrer at all. You'll never capture 100%—focus on trends.
Setting It Up in Your Analytics
- Create an "AI assistants" channel that matches the referrer domains and UTM sources above.
- Segment landing pages for that channel—which pages do AI assistants send people to?
- Build a conversion comparison: AI channel vs organic search vs paid.
- Filter session replays to AI-referred visits.
On Business and Enterprise plans, DeepSync's AI visibility feature isolates AI referral traffic and citations from answer engines automatically, alongside which AI crawlers read your site.
Measuring Quality, Not Just Volume
Track for AI-referred sessions:
- Engagement — pages per session, scroll depth.
- Conversion rate — goals and funnel completion (goals).
- Landing pages — which content earns citations.
- Bounce patterns — did the page match what the assistant promised?
What AI Visitors Do Differently
Watch 10–20 AI-referred sessions. Common patterns:
- They arrive deep in your site, not on the homepage.
- They look for one specific answer, often a price, feature, or comparison.
- They verify what the assistant said—if your page contradicts it, they leave.
- They are often further along in their decision.
Make cited pages conversion-ready
If an assistant keeps sending people to a blog post or docs page, add a clear next step there—a demo, signup, or relevant product page.
Key Takeaways
- AI assistants send real, often high-intent visitors.
- Identify them via referrer domains and UTM sources; some will look like direct.
- Measure conversion quality and landing pages, not just volume.
- Watch AI-referred sessions to make cited pages convert.
Conclusion
AI discovery is a new channel. Measure it now, while it's small, and you'll understand it when it isn't.
Frequently Asked Questions
Related articles
Is ChatGPT Recommending Your Brand? How to Check and Improve AI Visibility (GEO Guide)
11 min readAI Crawlers Explained — GPTBot, ClaudeBot, PerplexityBot, and Which to Allow
10 min readMarketing Attribution for Small Businesses — A Simple Setup That Actually Works
10 min readStay in the loop
Get the latest insights on product analytics and user behavior delivered to your inbox.



