The AI did the comparing before the click. The visitor arrives already recommended to.
What we measured
Google Analytics now separates visitors who arrive from AI assistants into their own channel, which makes a clean comparison possible for the first time: the same websites, the same months, AI-referred visitors against organic search and paid. We ran that comparison across our client properties, with their permission, covering tourism, travel and beauty businesses and a combined eighteen months of data.
The caveats first. These are hundreds of AI sessions, not tens of thousands, so we treat the numbers as a strong pattern rather than a law. And every figure below undercounts the real effect, because analytics only sees the visitor who clicks straight through from the AI. The one who hears a recommendation and then Googles the brand name lands in organic or direct, influenced all the same, invisible to attribution.
Our numbers
Across the client base, visitors arriving from AI assistants engage at 84% against 75% for organic search, and convert at 5.4% against 3.9%. Same sites, same measurement, different front door.

Client by client, the pattern sharpens with the size of the decision. Japan360, bespoke travel where an enquiry is a considered step, sees AI-referred visitors enquire at 12.5%: three times its organic search rate and six times its paid. En Beauty's AI visitors are the most engaged cohort on the entire site at 91%, and convert ahead of paid search. Pacific Cycle Tours' AI visitors spend an average of 5 minutes 24 seconds on site against roughly two minutes site-wide, the deepest attention of any channel.

The world's numbers agree, with a pattern worth knowing
The international studies land in a wide range, and the range itself is the insight. At the high-consideration end: Ahrefs published its own data showing AI referrals at 0.5% of traffic driving 12.1% of signups, roughly 23 times the conversion rate, and a 2026 benchmark of 312 B2B tech firms using combined analytics and CRM attribution measured AI-referred visitors converting at 14.2% against 2.8% for Google organic. Seer Interactive measured ChatGPT visitors converting at 15.9% against 1.76%. At the everyday-purchase end the lift is real but modest: a 15-month study across hundreds of e-commerce stores found 1.81% against 1.01%, and a separate year-long analysis of 94 stores found a 31% advantage. SE Ranking's behaviour data backs the mechanism, with AI visitors spending 68% more time on site. One analysis has even found AI traffic trailing organic on revenue per session, which is why we claim the pattern and not a magic multiplier.
The pattern: the bigger and more considered the decision, the bigger the AI advantage. Which describes most NZ service businesses, and matches exactly what our own data shows.
Why it works
The mechanism is unglamorous. When someone asks an AI assistant for a recommendation, the comparing happens before the click. The assistant has already weighed the options, checked the reviews, read the pages, and volunteered a name. So the visitor who arrives is not browsing ten blue links, they are following a recommendation they already trust, which is why they stay longer and convert more. Demand is moving to meet it: BrightLocal found ChatGPT use for local business recommendations grew from 6% to 45% of consumers in a year, and a 2026 multi-source analysis found 73% of B2B buyers now use AI tools in purchase research.
One thing the mechanism quietly requires: the whole chain has to hold. The recommendation, the page it cites, the experience when they arrive, all telling the same story, because the engines corroborate across everything they can read before they volunteer a name, and the visitor decides in seconds whether the page matches the promise. A great recommendation landing on a page that does not convert is a wasted gift, which is why the elements have to work in concert, not as separate projects.
What to do with this
Three moves, none of them expensive.
See your own number. The AI Assistant channel is native in GA4 now. Open your traffic acquisition report and compare its engagement and conversion against organic. If the channel barely exists for you, that is the finding: the recommendations are going to someone else.
Become recommendable. The answers are assembled from retrievable evidence: pages that answer buyer questions, reviews, consistent business details, a distinctive reason to name you. All of it is buildable. How to get your business recommended by AI →
Make the landing worth the recommendation. The visitors AI sends are the best-qualified traffic in the data, which makes conversion the multiplier on everything. What conversion rate optimisation involves →
And for the other half of this story, where the search demand went and why paid auctions are crowding, the companion piece: Why NZ businesses are suddenly buying Google Ads →
FAQ
Do AI referrals really convert better than Google traffic?
Across studies and our own client data, yes, with the advantage ranging from roughly 30% in e-commerce to several multiples for considered purchases and B2B. The common thread is pre-qualification: the AI compares options before the click, so the visitor arrives already leaning in.
How do I see AI traffic in my own analytics?
GA4 now groups it natively: open Reports, then Traffic acquisition, and look for the AI Assistant channel group. Compare its engagement rate, time on site and conversions against organic search. If the row is empty or tiny, your business is not yet part of the answers your buyers are getting.
Why do AI visitors convert so well?
Because the recommendation carries trust the click inherits. The assistant has already done the comparing, so the visitor is not evaluating a list, they are confirming a choice. The effect grows with the size of the decision, which is why travel, professional services and B2B see the largest lifts.
Source: client analytics measured in each client's own GA4 property, analysed November 2026. External figures: Ahrefs, Seer Interactive, ThoughtMetric, SE Ranking, BrightLocal, and a 2026 benchmark of 312 B2B technology firms.


