Here’s the conversation that tends to happen in most organisations once AI search comes up in a strategy meeting: someone asks what it’s actually worth, someone else points to AI referral sessions in GA4, and then someone, usually in finance, asks why that number is so small.
The answer is that observed AI referral traffic is the floor of AI’s commercial contribution, not the ceiling. And if you’re only reporting that number, you’re systematically understating the impact.
This post is about how to build a more complete, credible picture, one that accounts for the AI influence you can see and the influence you can’t, without making claims you can’t support.
The Four Confidence Layers
When you’re reporting on what AI search is actually doing for your business, don’t try to squeeze everything into one number. Instead, think of it as four separate buckets, each with a different level of reliability:
What you can see directly: This is traffic that arrives on your site from an AI platform and gets recorded in your analytics. Someone clicks a link in ChatGPT, lands on your site, you can see it. Solid data, but it only tells part of the story, plenty of people see your brand in an AI answer and never click, they just go and search for you later.
Clues from your own data: This is where you look for signs that AI is influencing people even when you can’t prove it directly. Things like: are more people searching for Koozai by name than usual? Is traffic growing to specific pages without any campaign pushing people there? Are more people ticking “AI assistant” when you ask how they found you on your contact form? None of these definitively prove AI drove it, but together they paint a picture.
What external tools tell you: Tools like Similarweb can estimate how much AI traffic you and your competitors are getting across the web. It’s not your own data so it’s less reliable, but it’s the only way to see how you’re doing compared to others, and to understand which types of questions are sending people to your pages in the first place.
Your best educated guess: This is where you take everything above and work out a rough commercial number for planning purposes. Something like: “We think AI influenced roughly 30% of our branded search growth this quarter. Based on how that traffic normally converts, that’s probably worth somewhere between £X and £Y in pipeline.” Be upfront that it’s an estimate with assumptions behind it; never present it as hard revenue or it’ll fall apart the moment someone asks how you got there.
Each layer answers a different question. Keep them separate in your reporting, label the confidence, and never present a modelled estimate as proof.
Setting Up GA4 to Catch What It Can
First, check whether you already have an AI channel in GA4
Go to Reports, then Acquisition, then Traffic Acquisition, and set the primary dimension to Session default channel group. If you’ve had AI traffic since May 2026, you should see an AI Assistant row appearing automatically. If it’s there, that’s your starting point.
But don’t rely on it alone. The built-in channel misses Perplexity entirely, and a significant chunk of AI-driven visits still land in Direct or Referral because the referrer information gets lost along the way. To catch more of it, you need to set up a custom channel group as well.
Setting up a custom AI channel group
Go to Admin, then Data display, then Channel groups. Create a new group, add a channel called something like “AI Search”, and set it to capture traffic from the main platforms by their domain names: chatgpt.com, perplexity.ai, claude.ai, gemini.google.com, copilot.microsoft.com, and any others showing up in your referral data. One thing to get right: custom channel groups apply from the date you create them and do not backfill historical data, so set this up as soon as possible. Also make sure your AI channel sits above the general Referral channel in the list as GA4 assigns traffic to the first matching rule it finds, so if Referral comes first, your AI traffic gets absorbed into it before the AI rule fires.
What about Google’s own AI features?
This is still a blind spot in GA4. Traffic from AI Overviews and AI Mode doesn’t appear as a separate source, Google’s own AI surfaces stay invisible in GA4 because Google doesn’t separately attribute them. The new Search Console reports help here, because they’ll show you impressions inside those features even if GA4 can’t track the clicks cleanly. Use both together.
Even with all of this set up properly, you’re still only seeing part of the picture. The sessions you can track directly are the floor of AI’s contribution, plenty of people see you in an AI answer, don’t click, and come back later through branded search or direct. That’s why the proxy signals covered earlier in this post matter just as much as the traffic data itself.
The Survey Question Worth Adding Today
There’s one proxy signal that’s disproportionately useful relative to how easy it is to set up: a single survey question in your signup or post-purchase flow.
Something like: “Before signing up, did you come across us in an AI assistant or AI search experience, such as ChatGPT, Perplexity, Claude, Gemini, or Google’s AI features?”
Yes / No / Not sure. Optional. One question, no follow-up.
The insight this gives you is significant. Users who arrive via branded search but answer “yes” are the invisible AI influence; they wouldn’t have searched for you without seeing you in an AI answer first. Users who arrive via direct traffic and answer “yes” are the mobile-to-AI copy-paste cohort that GA4 is entirely blind to.
A rising “yes” rate among branded and direct visitors is one of the strongest first-party signals you have that AI is driving real commercial impact beyond what your analytics can see.
How to Report This to Leadership
Nobody’s asking you to prove AI search is responsible for every new lead. The goal is simply to put together an honest monthly picture that helps you make better decisions about where to invest time and budget.
Here’s an example of what that could look like in practice for a digital marketing agency:
What GA4 is showing us directly
450 sessions arrived from AI platforms this month. Those visitors spent significantly longer on the site than our average organic visitor and were more likely to look at our case studies and service pages. Small number, but decent quality.
What our own data suggests
Searches for our agency name are up noticeably compared to the same period last quarter, without any paid activity or PR push that would explain it. Traffic to our SEO and digital PR service pages has also grown, despite us not running any campaigns pointing specifically at them. That pattern is consistent with people seeing us mentioned somewhere and then coming to check us out. AI answers are the most likely explanation.
What third-party tools are showing
Tools like Similarweb suggest the actual number of AI-influenced visits is probably higher than the 450 GA4 is recording, because a lot of AI-driven traffic loses its referrer data before it hits your analytics. On the competitive side, a rival agency appears to be getting a larger share of AI-referred traffic in our category, which tells us there’s ground to make up.
Our planning estimate
Taking the branded search growth above our normal baseline, and applying a conservative assumption that around a quarter of it is linked to AI influence, cross-checked against the number of new enquiries who told us they found us via an AI tool, we’d estimate AI search contributed somewhere in the region of £8,000 to £14,000 of influenced pipeline this quarter. The assumptions behind that are documented separately.
That last figure is a planning number, not something to drop into a client report as proof. The minute you present an estimate as hard revenue, someone will ask exactly how you got there and if the answer involves assumptions, you’ve undermined everything else on the page. Keep it clearly labelled as an estimate, show the working, and it becomes a genuinely useful tool for deciding whether to invest more in this area.
Three Patterns to Watch For
As you build up data across these layers, a few common patterns tend to emerge:
- Scenario 1: Hidden success. AI referral sessions are flat, but branded search is up significantly, direct traffic is growing, and your survey “yes” rate is rising. Reading: the visibility is working. Users are seeing you in AI answers and coming through other channels. Don’t cut investment because the observed channel looks quiet.
- Scenario 2: Traffic without fit. AI referral sessions are up, but conversion rate from AI traffic is below your organic benchmark. Reading: you’re getting cited for prompts that don’t match your most relevant pages. Check the prompts driving AI traffic and update the landing pages they’re pointing to.
- Scenario 3: Clean case. AI sessions up, branded search up, AI-assisted conversions visible, survey signal rising, competitive AI share growing. Reading: multiple independent signals all pointing in the same direction. Scale investment, expand prompt coverage, keep validating.
Pulling It All Together
If you’ve followed this series from the start, you’ll now have a framework that covers the full picture:
- Presence tells you whether and how you appear in AI search
- Readiness tells you why your visibility looks the way it does
- Business Impact tells you whether that visibility is creating commercial value
Running these three layers together turns what could be three disconnected audits into a single, connected diagnosis. You’ll know which lever to pull next; whether that’s fixing a technical access issue, building corroboration in key third-party sources, improving how you’re described on comparison pages, or simply reporting the impact you’re already generating more clearly.
AI search measurement isn’t about more dashboards. It’s about connecting what’s happening, why it’s happening, and what it’s worth, and being willing to act on what that tells you.
This series draws on frameworks developed by SEO consultant Aleyda Solis, including her 3-layer AI search measurement model and AI search prompt library guide. We’d recommend reading her original work for the full technical detail.
Leave a Reply