Resources
GuideJul 2, 20267 min read

AI sends you 50 visits. Google sends 5,000. Watch the 50.

AI referral traffic looks like a rounding error in every dashboard, and it converts almost like paid search. The size of the line is hiding what the line is worth.

By Yurii Shevchyk · Founder, AnswerPeek

Every time I show someone their AI referral numbers, the reaction is the same: that's it? Fifty visits against five thousand from Google looks like a curiosity, not a channel. Publishers make the same point louder — major outlets report less than 1% of referral traffic arriving from AI platforms, despite being cited constantly.

Then you segment by conversion instead of volume and the picture inverts. The 2026 benchmarks are remarkably consistent: ChatGPT referrals convert around 7.1% against roughly 2–3% for organic search — nearly the performance of paid search, at a cost per click of zero. Goodie's panel adds the engagement side: AI-referred visitors average 58.5 seconds of engaged time, about 30% longer than Google organic. Small line. Dense line.

Why the quality gap is structural, not luck

A Google click and a ChatGPT click are different moments in a decision. The searcher is starting research: ten blue links, tabs everywhere, high bounce. The ChatGPT visitor asked for a recommendation, read a paragraph that named two or three brands with reasons, and clicked one. The comparison shopping happened inside the answer. What lands on your site is the end of a funnel you never saw — which is exactly why the volume is small and the intent is not.

Google sends you an audience. AI answers send you a decision, nearly made.

The mention iceberg under the click

The referral line also undercounts on purpose, because a click is the rare case. Most of the time the model names a brand, the buyer reads it, and later types the name straight into the address bar or search box — direct traffic and branded search, wearing no AI label at all. The 50 visits you can attribute sit under an unknown multiple of influenced visits you cannot. Treat the visible line as a floor, never as the size of the channel.

It is no longer just ChatGPT, either

The same Goodie panel shows the engine mix moving fast: ChatGPT's share of AI referrals fell from 89% to about 63% in under a year, while Claude climbed from 1.4% to 18.5% — and Claude punches far above its weight in B2B, sending 18% of referrals from a small fraction of visits. If your buyers are technical or B2B, the fastest-growing slice of this channel may not be the engine you have been checking.

Make the 50 visible, then grow it

  • Build one AI-referral segment: chatgpt.com, perplexity.ai, claude.ai, copilot.microsoft.com, gemini.google.com. Report it beside organic, with conversion rate, not sessions, as the headline number.
  • Watch branded search and direct alongside it. A rising AI channel usually drags both — that correlation is your evidence for the invisible mentions.
  • Look at the landing pages AI actually sends people to, and make those pages work for a visitor with one sentence of context.
  • Run your buying prompts on a schedule. Mentions are the input; referrals are the lagging output. Improving the answers is how you grow the line.

For agencies, this is the cleanest report you can put in front of a client this year: a channel small enough to explain in one slide, converting near paid-search rates, with a fix list attached. The brands that dismiss the 50 visits are the same ones that dismissed the first 50 visits from Google in 1999. The line only stays small until it doesn't.

Questions, answered

How much traffic do AI engines actually send?
Very little in absolute terms — even heavily-cited publishers report under 1% of referrals from AI platforms in 2026. The channel's case is quality, not quantity: benchmark conversion rates around 7.1% for ChatGPT referrals versus roughly 2–3% for organic search.
Why does AI referral traffic convert so much better than organic?
Because the model pre-qualifies the visitor. They asked for a recommendation, read an answer that named and justified a shortlist, and clicked a brand the AI endorsed. The research and comparison stage finished inside the answer, so the site receives a nearly-made decision.
Which AI engines should I track referrals from?
Segment chatgpt.com, perplexity.ai, claude.ai, copilot.microsoft.com and gemini.google.com together as one channel. Watch the mix: ChatGPT still leads, but its share of AI referrals dropped from about 89% to 63% within a year while Claude grew to 18.5% — with a notable skew toward B2B audiences.
Does AI visibility show up anywhere besides referral traffic?
Mostly elsewhere, in fact. Many buyers read a recommendation and later arrive as direct traffic or branded search, with no AI referrer attached. Rising branded search alongside a growing AI segment is the usual signature of mentions doing their work invisibly.

See where you stand in AI answers.

Keep reading

Guide

ChatGPT started linking to brands. Most teams haven't noticed.

In May ChatGPT quietly began placing clickable brand links inside its answers, and referral traffic jumped overnight. Here is what changed, and how to be on the right side of it.

7 min read