AI Traffic
AI traffic is website visits that come from AI assistants like ChatGPT and Perplexity, plus the crawler visits those engines make. Learn how to track both.
TL;DR
AI traffic is website traffic connected to AI assistants, and it comes in two forms that behave nothing alike. AI referral traffic is human visitors who clicked a link inside an answer from ChatGPT, Perplexity, Gemini, or Copilot. AI crawler traffic is the bots those companies send to read your pages, with no human attached.
Why AI traffic matters
Visits from AI answers arrive further along the buying process, because the assistant already did the comparison work before sending anyone your way.
- Intent quality: A visitor who clicked through an AI recommendation has already seen your brand described and vetted against alternatives.
- Attribution gaps: AI referrals often land in "direct" or "unassigned" buckets, so the channel looks smaller than it is until you tag it.
- Falling search clicks: As AI answers absorb informational queries, referral traffic from those answers replaces some of what organic search used to deliver.
- Infrastructure load: Crawler traffic from AI companies consumes server resources without ever converting, which belongs in a cost conversation.
How to track AI traffic
Find the referrers
In GA4, build a report filtered to referral sources containing chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com, and claude.ai.
Separate the crawlers
Filter server logs by user agent for GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and Applebot-Extended. These never appear in GA4 because they do not run JavaScript.
Watch for missing referrers
Some assistants strip the referrer header, so compare direct-traffic growth against your AI referral trend before concluding the channel is small.
Segment behavior
Compare session duration, pages per session, and conversion rate for AI referrals against organic search. The differences tell you whether the answers describe you accurately.
Set crawl rules
Decide per bot what you allow in robots.txt. Blocking a training crawler and blocking a retrieval crawler have very different consequences for visibility.
AI referral traffic vs. AI crawler traffic
| AI referral traffic | AI crawler traffic | |
|---|---|---|
| Who | Humans arriving from a link in an AI answer | Automated requests from AI company bots |
| Visible in | Analytics | Server logs only |
| Converts? | Yes, capable of converting | Never converts |
| What it tells you | A direct read on whether AI engines recommend you | Whether your pages can be cited at all |
Referral traffic only tells you about the answers that sent someone to you. It says nothing about the answers where a competitor got named instead, which is most of them. Findrix runs the questions your market asks across ChatGPT, Gemini, Perplexity and four other engines every week, then shows how often each one cites you against named competitors. Every gap comes with the fix already written: technical, content and off-site. The audit is free, takes about a minute, and requires no signup.
Metrics for AI traffic
- Referral sessions by engineSessions split by assistant, tracked weekly, since engines change their citation behavior at different times.
- Conversion rate by sourceAI referrals compared against organic and direct, which is the honest test of traffic quality.
- Crawl frequency per botHow often each AI crawler fetches your site, and which sections it favors.
- Crawl-to-citation lagThe gap between a crawler fetching an updated page and the engine repeating the new information.
- Assisted conversionsAI referrals often appear early in multi-touch paths, so last-click reporting understates them.
Why AI traffic numbers look small
Most sites see AI referrals as a low single-digit share of sessions, which leads teams to dismiss the channel. Two things distort that reading. First, undercounting: stripped referrer headers push an unknown share of AI visits into direct traffic, and few analytics setups are configured to catch them.
Every answer that names your brand shapes a buying decision whether or not anyone clicks, so referral sessions measure the leakage from AI answers rather than the influence of them.
Frequently asked questions
Is AI traffic good or bad for a website?
Referral traffic from AI answers converts well because those visitors arrive pre-qualified. Crawler traffic is a cost with an upside: bots that cannot fetch your pages cannot cite them. The sensible approach is to welcome retrieval crawlers and decide separately about training crawlers.
Why does AI traffic not show up in Google Analytics?
Two reasons. Crawlers never execute the JavaScript that GA4 depends on, so they are invisible there and appear only in server logs. And some assistants omit the referrer header on outbound links, which sends those human visits into direct traffic instead of a named source.
Should I block AI crawlers in robots.txt?
Treat the decision per bot. Blocking retrieval crawlers such as OAI-SearchBot or PerplexityBot removes you from the answers those engines generate. Blocking training crawlers such as GPTBot limits use of your content in model training with a smaller effect on live citations.