Use both GA4 and Google Search Console to measure AI search traffic, but use them for different jobs: GA4 is better for on-site behavior and conversions, while Search Console is better for search visibility, queries, and clicks. Google says traffic from AI features such as AI Overviews and AI Mode is included in Search Console reporting for web search, but it is not neatly split into a clean “AI Mode” report [source: Google].
TLDR: GA4 can help you spot possible Google AI Mode referrals, especially when the referrer contains signals such as udm=50, then connect those visits to engagement, leads, and revenue. Search Console shows the broader search picture, including clicks, impressions, queries, and landing pages, but it does not give you a simple AI Mode traffic filter. For example, a SaaS site might find 420 GA4 sessions from suspected AI Mode referrals in 30 days, with a 6.8% trial signup rate, while Search Console shows 18,000 impressions and 760 clicks for the same landing pages. Use GA4 to judge business value, and Search Console to judge search demand.
GA4 vs Search Console: the short version
GA4 and Google Search Console often get compared as if one should replace the other. That is the wrong frame. They answer different questions.
- GA4 answers: What did visitors do after they landed on my site?
- Search Console answers: How did my pages perform inside Google Search?
- GA4 is session and event focused. It tracks visits, engagement, conversions, revenue, and user paths.
- Search Console is search result focused. It tracks impressions, clicks, click through rate, queries, and average position.
For AI search traffic, this difference matters a lot. AI Mode can send fewer clicks than classic blue links, but those clicks may be highly qualified. GA4 helps you prove that. Search Console helps you see whether your content is being exposed in search at all.
What the GA4 Google AI Mode referrer can tell you
GA4 records where traffic came from through fields such as source, medium, session source medium, and page referrer. When someone clicks from Google AI Mode to your site, GA4 may classify it as google / organic. In some cases, the full referrer can include details that suggest the visit came from AI Mode.
The signal many analysts watch is udm=50, which has been associated with Google AI Mode search URLs. If your GA4 setup stores the full page referrer, you can build reports that filter Google referrers containing this parameter.
The catch is that this feels more like detective work than proper reporting. One missing query parameter, one privacy setting, or one browser policy can hide the clue. Expect to waste time checking whether GA4 preserved the full referrer or trimmed the useful part.
Still, GA4 gives you something Search Console cannot: post-click quality. You can compare suspected AI Mode sessions against normal organic sessions by metrics such as:
- Engagement rate
- Average engagement time
- Key event rate
- Lead form submissions
- Purchases or subscription starts
- Landing page conversion rate
That is where GA4 earns its place. If AI Mode sends only 2% of organic sessions but drives 9% of demo requests, you need to know that fast.
How to find possible AI Mode traffic in GA4
Start with an Explore report in GA4. Use dimensions such as Page referrer, Session source / medium, Landing page + query string, and Page path. Then filter for referrers from Google that include suspected AI Mode patterns.
A simple regex could look like this:
google\..*\/search.*udm=50
This is not perfect. Treat it as a working signal, not a guaranteed official label. If you use BigQuery export, you can inspect referrer strings with more control and avoid some of the UI pain. GA4 dimensions and traffic source rules can also behave differently depending on reporting scope, so check whether you are reading event level, session level, or user level data [source: Google Analytics Help].
Honestly, it feels like Google could save everyone hours by adding a built-in “AI Mode” dimension. Until that happens, your best GA4 setup is a custom exploration, a saved audience or segment, and a Looker Studio dashboard that separates suspected AI Mode from standard organic search.
What Search Console gives you instead
Google Search Console does not show on-site behavior. It will not tell you whether AI Mode visitors bought, subscribed, downloaded, or bounced after eight seconds.
But it gives you something GA4 cannot measure: impressions. That is vital for AI search. Many users may see your brand, product, or article cited in an AI answer and never click. GA4 sees nothing there. Search Console may still show impressions and clicks tied to the query and landing page, depending on how Google reports that surface.
Use Search Console to monitor:
- Queries that trigger your pages in Google Search
- Landing pages gaining or losing visibility
- CTR changes after AI answers appear more often
- Impression growth without matching click growth
- Pages with high impressions but weak clicks
This is where a common pattern appears. A page may gain impressions but lose CTR because the AI answer satisfies part of the user’s need. That does not always mean the page is failing. It may mean the content needs stronger original data, clearer product fit, or a better reason to click.
Where GA4 wins
GA4 wins when the question is commercial.
Ask GA4:
- Did suspected AI Mode visitors convert?
- Which landing pages turned AI search traffic into leads?
- Did these users read one page or continue deeper?
- Which content assisted paid signups later?
- Are AI search visitors more valuable than standard organic visitors?
For example, an ecommerce store might see only 300 suspected AI Mode sessions in a month. That sounds tiny. But if those users produce $4,800 in revenue and the average order value is 28% higher than standard organic traffic, the segment matters.
Where Search Console wins
Search Console wins when the question is visibility.
Ask Search Console:
- Which queries are growing?
- Which pages are being shown more often?
- Is CTR falling while impressions rise?
- Did a content update affect search demand?
- Are informational pages still earning clicks?
This is especially useful for content teams. If a how-to guide shows 40,000 impressions and only 220 clicks, it may be appearing in answer-heavy results. The fix is not always “write more.” Sometimes the fix is adding comparison tables, original screenshots, expert commentary, pricing details, or tools that AI summaries cannot fully replace.
The best measurement setup
The strongest approach is simple. Use Search Console for search exposure and GA4 for business impact. Then compare them by landing page.
- Export Search Console data by page, query, impressions, clicks, CTR, and position.
- Create a GA4 segment for suspected AI Mode referrals using Google referrer patterns.
- Compare landing pages across both tools.
- Flag mismatches, such as high impressions but low GA4 sessions.
- Report conversions separately for suspected AI Mode, standard organic, paid search, and referral.
A practical monthly report might include five columns: Search Console clicks, Search Console impressions, GA4 organic sessions, suspected AI Mode sessions, and key events. This makes the tradeoff visible. You can see whether AI search is creating awareness, traffic, or actual revenue.
Final recommendation
Do not choose between GA4 and Google Search Console. Pair them. GA4 is your truth for user behavior after the click. Search Console is your truth for how Google Search exposed your pages before the click.
For the GA4 Google AI Mode referrer, build a cautious custom report and label it clearly as suspected AI Mode traffic. Do not oversell it as perfect. Referrer data can break, vanish, or get grouped into normal organic search.
If you need one north star metric, use this: AI search assisted value by landing page. Match Search Console visibility with GA4 conversions. That gives you a cleaner read on whether AI search is costing you clicks, sending better visitors, or quietly changing which content deserves your next update.
