ThriveStack citedbyAI search attribution learning guideIntroductionLive demoBook a demoStart free
Pillar 3 · 6 lessons

Measurement:
referrals, GA4 and revenue by AI engine

Measurement turns AI answers into numbers you can report. Each engine tags or hides its visits differently, so you measure each one in its own way, then join visits to signups and revenue. Treat every figure as a floor.

What this pillar covers

Start with ChatGPT, the largest source of AI referrals. Then cover the part of Google you cannot isolate, Perplexity, and the smaller engines. Then set up GA4 so AI visits report as their own channel.

The last lesson joins the visit, the signup and the payment, so AI revenue shows next to sessions. For the tag and domain details, see lesson 4.1 on UTM tags and lesson 4.2 on referrer domains.

AI referral traffic and tracking, in numbers.

0.32%

of website traffic came from ChatGPT in May 2026, up from 0.23% in April

SE Ranking, 101,574 sites, Jul 9, 2026
8% vs 15%

of visits clicked a normal result with and without an AI summary

Pew Research Center, Mar 2025
May 13

2026: GA4 added an AI Assistant default channel

Google Analytics Help, via the ThriveStack engine guide
83%

higher revenue per visit for AI traffic on U.S. retail sites

Adobe, Sep 2026, published Oct 29, 2026

Four steps to measure AI traffic and revenue by engine.

  1. 1

    Filter GA4 to your top AI engine

    Find chatgpt.com in GA4 session source and add it to an AI channel group. Lesson 3.1 shows how.

  2. 2

    Set up a GA4 AI channel group once

    Use the AI Assistant channel, add a custom group for the rest, and test one tagged click. Lesson 3.5 has the steps.

  3. 3

    Watch direct traffic next to your AI channel

    A fall in tagged visits with a rise in direct points to a lost tag, not lost interest.

  4. 4

    Join AI visits to signups and revenue

    Carry the source into your CRM and report AI-sourced revenue as a floor. Lesson 3.6 has the join.

Common questions about measuring AI traffic by engine, answered.

Which AI engine should I measure first?

ChatGPT. SE Ranking found it sent 74.78% of AI referral traffic across 101,574 sites (Jun 18, 2026). Perplexity is easy to filter too.

Can I measure Google AI Mode and AI Overviews?

Not as a clean channel. Visits land in google / organic. Use branded search lift, a prompt panel and a survey question, and label the result as an estimate.

Why do my AI numbers look low?

Some engines strip the referrer, so visits land in direct. Treat AI-sourced figures as a floor and add a survey question to catch the rest.

Sources for this pillar

  1. SE Ranking: ChatGPT referral traffic, May 2026 (Jul 9, 2026)
  2. Pew Research Center: Google users and AI summaries (Jul 22, 2025)
  3. Adobe: AI traffic drives 83% higher revenue per visit (Oct 29, 2026)
  4. Cloudflare: the crawl-to-click gap (Aug 29, 2025)
  5. ThriveStack: which AI engines send utm_source

Figures were checked against each page in Nov 2026. Vendor studies are marked by their sample.

MapHover to open