The fastest-growing traffic source is the one nobody tags
In the first quarter of 2026, AI-referred traffic to US retail sites grew 393% year over year, as Fast Company reported in its coverage of the agentic commerce race. And these are not idle clicks. Visitors who arrive from AI assistants outperform traditional traffic on conversion rate, revenue per visit, time on site, and engagement.
The logic is simple once you see it. Someone who asked ChatGPT or Perplexity a question, got an answer that cited you, and then clicked through has already done the comparison work inside the chat. They land on your site warm, with context, often ready to act.
Yet in most analytics setups this segment does not exist. It gets lumped into a generic "referral" bucket, or worse, into "direct," and the most valuable visitors you have disappear into an average.
Why your reports are blind to it
Three habits create the blindness:
- Unsplit referrers. Traffic from chatgpt.com, perplexity.ai, gemini and their siblings arrives with a referrer you could segment, but the default channel groupings in most tools do not break it out. You have to define the segment yourself.
- UTM indiscipline. UTMs only exist on links you place yourself. Links generated inside an AI answer carry none, so a reporting culture built entirely around campaign tags is structurally unable to see this channel.
- Blended averages. If AI-referred visits sit inside the general referral pool, their higher conversion and revenue per visit get diluted until the signal vanishes.
The signal beyond the click
Measurement here is not only referrer parsing. According to Search Engine Journal, when an AI Overview appears on a branded query, click-through actually rises by 18%: clicks redistribute toward the brands the answer cites. Visitors coming through AI Overviews arrive with more context and higher intent.
That means two extra things belong on your dashboard: your branded search volume over time, and which of your pages actually get cited in AI answers. A page that gets quoted is doing sales work you are currently not crediting.
A minimal measurement setup
You do not need a data team for this. Four steps:
- Define an "AI referral" segment in your analytics tool that matches the known assistant domains, and keep the list maintained as new assistants appear.
- Report that segment's conversion and revenue separately, next to organic and paid, not blended into them.
- Watch citations, not just visits. Check periodically which pages assistants quote when asked about your category, and track branded query volume as a trailing indicator.
- Start light. A lightweight tool such as Vercel Analytics is enough to begin; it is what we run on our own site at Clodron. Sophistication can come after the segment exists.
What to do this week
Create the AI referral segment today, backfill as far as your data allows, and put its conversion rate next to your other channels in the next reporting cycle. If the pattern in the retail data holds for you, this will be your best-converting segment, and every content and product decision that increases your citability suddenly has a number attached to it. You cannot grow a channel you refuse to see.