Cleotic

AI Referrals

AI referrals are real human visitors who clicked a link in an AI chatbot response (like ChatGPT, Claude, or Perplexity) and landed on your website. While crawler evidence in Sources -> AI traffic tells you which bots are indexing your content, AI referrals tell you something more directly valuable: whether your AI visibility is actually driving traffic to your site.

How it works

Cleotic detects AI referrals automatically using the same server-side tracking you've already set up for crawler tracking. No additional setup is needed.

For each request that hits your site, Cleotic inspects:

  • The HTTP Referer header -- when a user clicks a link inside an AI chat, their browser typically includes the AI assistant's domain (e.g., https://chatgpt.com/) as the referrer.
  • The utm_source query parameter -- some AI tools tag outbound links with ?utm_source=chatgpt.com instead of (or in addition to) sending a referrer header.

If either matches a known AI chatbot, the visit is recorded as an AI referral.

Detected AI sources

Cleotic currently detects referrals from:

  • ChatGPT (OpenAI) -- chatgpt.com, chat.openai.com
  • Claude (Anthropic) -- claude.ai
  • Perplexity -- perplexity.ai
  • Gemini (Google) -- gemini.google.com, bard.google.com
  • Copilot (Microsoft) -- copilot.microsoft.com, bing.com/chat
  • Meta AI -- meta.ai
  • You.com -- you.com
  • Le Chat (Mistral) -- chat.mistral.ai
  • Grok (xAI) -- grok.com, x.com/i/grok
  • Character.AI -- character.ai
  • DuckDuckGo AI -- duckduckgo.com/aichat

What you'll see

The Sources -> AI traffic explorer includes referral evidence alongside crawler evidence. Set the Traffic filter to Referrals only when you want to focus on human visits from AI assistants. The current explorer shows:

Stats

  • Referrals -- the count of human visitors arriving from AI chatbots in the selected range
  • Pages -- how many distinct pages received AI traffic
  • AI systems -- which providers or sources sent traffic

Page-level referrals

The AI traffic table shows pages that received referrals, the AI sources involved, latest seen time, and trend. Click a page row to inspect recent matching evidence. Use the provider/source filters to narrow the table to a specific AI assistant; the counts and ranking adjust accordingly.

Why it matters

Crawler activity is a leading indicator -- it tells you AI systems can see your content. AI referrals are a lagging indicator -- they tell you AI systems are actively recommending you to real users.

Use this data to:

  • Validate your AI visibility strategy. If your visibility scores are climbing but referrals stay flat, AI tools may know about you without recommending you.
  • Identify your strongest AI evangelists. Knowing which AI assistants drive the most traffic helps you focus optimisation efforts.
  • Find your highest-performing landing pages. Pages that win AI referrals are already doing something right -- consider creating more in the same vein.
  • Measure campaign impact. Launched a new product page or content piece? Watch referrals to see if AI tools start surfacing it.

Notes and limitations

  • Some AI assistants strip referrer headers for privacy. We catch what we can.
  • We exclude self-referrals (e.g., a user clicking from one of your own pages to another) so the numbers reflect external AI traffic only.
  • Server-side tracking gives the most complete referral picture because it can inspect request headers before page JavaScript runs. JavaScript-capable AI browsers may also be recorded through the tracking pixel.