Every marketing manager I talk to right now is asking some version of the same question: “I hear AI search matters, but how do I actually prove it’s working?”
Fair question. Traditional SEO had Google Search Console, clean GA4 referral data, and keyword rankings you could watch. AI search is messier — and if you don’t know where to look, it can feel like nothing’s happening at all.
Good news: you can measure this. You just have to know where to look.
If you haven’t read the pillar on this yet, start here: SEO Isn’t Dead. It’s Entered a New Era. — it covers how AI search actually works and what to build for it. This post picks up where that one leaves off.
Here are the four things every marketing manager should be tracking right now.
1. AI Referral Traffic in GA4
The easiest place to start. Some AI platforms now pass referral data to your analytics — so you can see visits coming from ChatGPT, Perplexity, and others.
Where to look: GA4 → Reports → Acquisition → Traffic Acquisition.
Watch for these sources:
chatgpt.com / referral— visits from ChatGPT citationsperplexity.ai / referral— Perplexity citationsbing.com / organic— Bing search traffic (remember, ChatGPT runs on Bing)google.com / organic— traditional search, but now includes AI Mode/AI Overview clicks
OpenAI confirms in its publisher docs that ChatGPT automatically tags referral URLs with utm_source=chatgpt.com — so that traffic is clearly attributable when it arrives.
Translation: if you’re not seeing any of these sources in your GA4, either you’re not being cited yet, or your tracking isn’t catching them. Both are fixable.
2. Create an “AI Landing Page” Segment
Once you’ve built pages optimized for AI answers — your content hub pages, FAQ pages, the chunk-friendly resources — you want to isolate how they perform separately from the rest of your site.
Where to look: GA4 → Explore → new Exploration → segment filtered to those specific URLs.
Track, for that segment only:
- Total sessions
- Referral source mix
- Engagement rate
- New vs returning users
This tells you whether your AI-targeted content is actually doing the job, and gives you clean before/after numbers when you optimize further. Before you change anything on a page, take a screenshot of the numbers. Two months later, compare.
3. Branded Search Growth
This is the most underrated metric in AI visibility.
When AI models cite you, people often don’t click the link — they remember your name and Google you directly later. That shows up as growth in branded search queries over time.
Where to look: Google Search Console → Performance → filter queries for your brand name.
If your branded search volume is trending up while your paid advertising and PR haven’t changed, AI citations are probably driving awareness. That’s a compounding signal — every month of citations builds more brand recognition, which builds more branded searches, which builds more trust signals for future citations.
Set a baseline this month. Check it quarterly.
4. Direct Traffic Pattern Analysis
The messiest metric, but often the most telling.
When someone gets your URL or brand name from ChatGPT or Perplexity without clicking a link (because the AI mentioned your site by name), they often type it directly into their browser. That shows up in GA4 as direct traffic.
Where to look: GA4 → Reports → Acquisition → Traffic Acquisition → Direct channel.
Watch for:
- New users (not returning — that’s brand awareness growth, not loyalty)
- Specific landing pages — if people are hitting
/services/or/about/directly, something is sending them there - Geography shifts — AI citations can drive traffic from unexpected markets
It’s not perfect attribution, but a rising tide of direct new-user traffic to your money pages is usually AI doing its job.
What these four metrics tell you together
None of these signals alone is perfect. Together, they paint a clear picture:
- AI referral traffic → are AI platforms sending you traffic at all?
- AI landing page segment → is your optimized content pulling its weight?
- Branded search growth → are citations building awareness over time?
- Direct traffic patterns → is your brand being remembered outside clickable links?
If all four are moving up, your AI search optimization is working. If one is flat, you know where to focus next.
If you’re starting from scratch
If you haven’t set any of this up yet, here’s the first-week checklist:
- Confirm GA4 is installed correctly and connected to Google Search Console
- Create an “AI Landing Page” segment in Explorations for your key pages
- Screenshot your branded search volume this month as a baseline
- Pull a 90-day direct traffic trend report and save it
- Set a calendar reminder to repeat the same checks in 90 days
Three months of tracking gives you a trend. Six months tells you what’s actually working.
The bottom line
Measuring AI search ROI isn’t impossible — it just requires looking in the right places and accepting that attribution is messier than it used to be.
The businesses winning at this aren’t waiting for perfect measurement. They’re tracking the signals they can see, building the content that gets cited, and letting the brand awareness flywheel do its work.
If you want the full picture of why these metrics matter and how AI search is actually reshaping SEO, read the pillar post: SEO Isn’t Dead. It’s Entered a New Era.
Want help setting up AI search measurement on your site? Book a 15-minute Zoom and I’ll walk you through what to track and how.
