THE SHORT ANSWER
Track identifiable assistant referrals through analytics source and medium, landing pages and tagged links where platforms provide them. Connect those sessions to useful actions and assisted conversions. Add traditional search performance, branded demand and a documented sample of citations or mentions. Referrer loss, app behavior, copied URLs and zero-click answers create gaps, so do not claim complete AI attribution.
Build the measurement stack from observable events
| Layer | Possible measure | Main limitation |
|---|---|---|
| Referral | Session source/medium and referrer | Apps or browsers may omit or alter it |
| Landing | Pages receiving identifiable visits | Does not show unseen answers |
| Outcome | Enquiry, signup, sale or accepted task | Attribution may miss earlier influence |
| Search | Queries, impressions, clicks and feature reporting | Platform scope and aggregation differ |
| Brand | Branded searches, direct visits and recall research | Cause is difficult to isolate |
| Citation sample | Recorded references for defined prompts | Not population-wide visibility |
Use source and medium without over-reading them
GA4 defines session source, medium, channel and landing-page dimensions that can describe observable visits. OpenAI currently says ChatGPT search referral URLs include `utm_source=chatgpt.com`. A reporting rule can group known sources, but the list needs maintenance and does not recover visits with missing referral information.
Inspect landing-page engagement and business outcomes rather than treating every assistant session as equal. Keep raw source values available so a grouping rule can be audited.
Evidence & context: Google Analytics · OpenAI Help Center
Treat sampled answers as observations
- Choose questions tied to actual audience decisions.
- Record prompt, date, interface, locale and account state where relevant.
- Capture visible links, citations and brand descriptions.
- Check whether the information is accurate and current.
- Repeat enough to see variability, but do not call the sample market share.
Google's reporting for its own AI Search features follows its current platform definitions. It should not be combined with assistant prompt checks as if both used the same denominator.
Evidence & context: Google Search Central · Google Search Central
Connect visibility to contribution and value
- Eligible and indexed information.
- Visibility in relevant search and sampled answer contexts.
- Accurate representation and useful references.
- Qualified visits and landing-page behavior.
- Conversions, assisted outcomes and customer quality.
- Cost of content, measurement and maintenance.
Sources & further reading
- API dimensions and metrics
Google Analytics. Official GA4 reporting definitions checked 13 September 2026, including session source, medium, referral and landing-page dimensions. Attribution remains limited to observable interactions.
- AI features and your website
Google Search Central. Official documentation for AI Overviews and AI Mode checked 13 September 2026. Product behavior, availability and reporting can change; eligibility never guarantees crawling, indexing or appearance.
- Publishers and Developers — FAQ
OpenAI Help Center. Official publisher guidance checked 13 September 2026. It distinguishes OAI-SearchBot discovery from GPTBot training controls and documents ChatGPT referral tagging; crawler roles can change.
- AI features and your website
Google Search Central. Official discovery and Search Console reporting guidance, checked 11 September 2026. Visibility and indexing are not guaranteed.
Examples and exercises are illustrative unless attributed to a source. No independent expert review is claimed.
A correction, a counterexample or an experience worth sharing?
Join the conversation ↗