Attributing Visits from AI Platforms

Traffic from AI assistants is growing, but in analytics it tends to be hidden or misclassified – it often merges with direct visits or generic referral. Without correct attribution you underestimate a channel that is gaining importance and deprive yourself of the argument for investing in AIEO.

How to recognize AI traffic

• Referral from domains such as chat.openai.com, perplexity.ai, gemini.google.com and others.
• Links that AI takes over from cited sources – the user clicks from the answer through to your site.
• UTM parameters everywhere you can add them (for example in links you place yourself).

Why it is not trivial

Part of AI traffic is attributed incorrectly. When someone reads an answer in ChatGPT and then manually searches for your brand or types it directly into the browser, it shows up as organic or direct traffic – not as AI. Precise attribution will therefore never be one hundred percent; the point is to capture the trend, not every single visit.

Create a segment

In analytics, set up a separate segment for known AI referrals (by source domains) and track its volume, on-site behavior and conversions separately. Only when you see AI traffic as its own channel can you assess its real contribution and compare it with the others.

Evaluate in context

AI traffic tends to be smaller in volume but above average in quality – the user arrives with a specific intent, often closer to a decision. So judge it by conversions and value, not by the number of visits. A small but well-converting channel can be more valuable than a large but shallow one.