Technical:
UTMs, referrers, logs, CRM and privacy
This pillar has 7 lessons. Each one answers the questions buyers ask on one topic, with a visual, the numbers behind it and a checklist you can use this week.
UTM tags and tracking
2 of 18AI engines seen adding a utm_source tagChatGPT UTM source attribution
Lesson 4.2 · 5 questionsSpot AI traffic by referrer
~91%of Gemini app mobile traffic is darkHow do AI assistants pass referrers?
Lesson 4.3 · 10 questionsServer logs and crawler signals
~80%of AI crawling was for model trainingAI crawler server log analytics
Lesson 4.4 · 10 questionsMatch AI visits to CRM contacts
7sources the average B2B buyer used in a recent purchaseConnect AI referrals to HubSpot contacts
Lesson 4.5 · 10 questionsData pipelines for AI attribution
DiagramA visual walks through this lesson.AI traffic in Looker Studio
Lesson 4.6 · 10 questionsAgentic and AI browser traffic
Directwhere ChatGPT Atlas visits often land, or (not set)How to track ChatGPT Atlas traffic
Lesson 4.7 · 10 questionsPrivacy, consent and compliance
DiagramA visual walks through this lesson.Is AI search attribution GDPR compliant
What this pillar covers
Start with the signals an AI visit can carry. Lesson 4.1 covers the utm_source tag that a few engines add, and lesson 4.2 covers the referrer domain that most engines send. Lesson 4.3 adds server logs, which show the AI crawlers that never run your analytics script.
Then connect the pieces. Lesson 4.4 carries the AI source from the first visit into your CRM, lesson 4.5 joins GA4, logs, CRM and billing into one report, lesson 4.6 separates people in AI browsers from AI agents, and lesson 4.7 covers privacy and consent for each signal.
AI tags, referrers and crawlers, in four numbers.
AI engines seen adding a utm_source tag
ThriveStack engine guide, Oct 16, 2026of Gemini app mobile traffic is dark
Searchpilot, via ThriveStack engine guide, Oct 2026of AI crawling was for model training
Cloudflare, Aug 29, 2025the ChatGPT tag was gone, Oct 9 to Oct 15, 2026
ThriveStack engine guideFour steps to capture every AI signal on your site.
- 1
Tag every link you control
Add utm_source, utm_medium and utm_campaign to the links you place, so those visits never depend on an AI engine adding a tag. Lesson 4.1 shows the format.
- 2
Add an AI referrer channel in GA4
Match the known AI referrer domains with one regex in a channel group, and date every rule you add. Lesson 4.2 has the domain list.
- 3
Read AI crawler hits in your server logs
Pull AI crawler hits from your host or CDN logs, so you see which pages AI engines read. Lesson 4.3 shows what to look for.
- 4
Store the AI source on the contact
Capture the source on the first visit, save it on the contact at signup, then copy it to the deal. Lesson 4.4 lists the fields.
Common questions about AI tags, logs and CRM data, answered.
Do AI engines add UTM tags to their links?
A few do. Only ChatGPT and Gemini are seen adding utm_source, and the ChatGPT tag was gone for about six days in October 2026. Treat any tag an engine adds as a bonus, and use referrers and server logs as well. Lesson 4.1 has the details.
Why do I need server logs if I already have GA4?
AI crawlers do not run JavaScript, so GA4 never sees them. Only server logs show which pages AI engines read, and how often. Lesson 4.3 explains how to read them.
Is AI search attribution allowed under GDPR?
It can be, if you treat it like any other analytics. Server logs hold IP addresses, which count as personal data under GDPR, so shorten or hash them and set a retention period. A source cookie needs consent in many regions. Lesson 4.7 walks through each signal. Check the details with your own counsel.