AI search referrals convert better but the traffic drop is steeper
eMarketer's latest data reveals a small but high converting AI referral segment that most retailers are not yet tracking properly.
Mariano De Vitto · September 2026
New data from eMarketer is forcing marketers to rethink how they measure the payoff of AI search. Visits arriving from AI search platforms convert at notably higher rates than traditional referral traffic, yet these AI referrals still make up less than one percent of total referral visits for most retailers. The finding, detailed in eMarketer's analysis of how retailers are adapting to AI as a referral channel, signals a shift that marketers cannot ignore even though the volume remains small for now.
eMarketer, the research firm that has tracked digital commerce trends for years, released this latest look at AI referral traffic as retailers scramble to understand where tools like ChatGPT and Perplexity fit into their broader acquisition mix. The numbers are small today, but the behavioral pattern behind them is what has caught analysts' attention.
Why AI shoppers convert faster
The report highlights that shoppers who click through from tools like ChatGPT or Perplexity tend to arrive further along in their purchase decision, having already used the AI assistant to compare products or narrow down options. Rather than starting from a blank search box, these users have effectively completed much of their research inside the AI interface before they ever reach a retailer's site. This behavior explains why conversion rates from these sessions outperform legacy search and even some paid channels, according to eMarketer's findings.
That distinction matters for how retailers organize their analytics. Retailers cited in the report are beginning to treat AI referral traffic as a distinct segment worth tracking separately from organic and paid search, rather than lumping it into general referral metrics. Grouping AI clicks with generic referral traffic would mask both the small volume and the outsized conversion quality, making it harder to justify further investment in AI visibility.
The gap between visibility and clicks
The catch, per eMarketer, is that the sharp drop in AI referral traffic between initial visibility and an actual click is steeper than what marketers see with traditional search engines. A brand can appear prominently in an AI generated answer without that visibility translating into a session, since users often get their answer directly from the AI interface and never click through to a website at all. This is a structural difference from classic search, where ranking well on a results page has historically correlated fairly reliably with earning a share of clicks.
For marketers accustomed to measuring success through click through rate and session volume, this creates a blind spot. A brand could be winning the recommendation inside ChatGPT or Perplexity for weeks without a single referral dollar showing up in a standard analytics dashboard, simply because the platform answered the user's question directly.
What marketers should do next
Why it matters: marketers need to stop measuring AI search purely by session counts and start tracking visibility within AI generated answers themselves. Since less than one percent of referral traffic currently comes from AI platforms, teams that wait for volume to justify investment risk missing the higher intent, higher converting audience that is already there.
Building visibility tracking into SEO reporting now positions brands ahead of competitors still focused solely on traditional click through metrics. The retailers already adapting, according to eMarketer, are the ones treating AI referral behavior as a leading indicator rather than a rounding error, and that early instinct is likely to pay off as AI assisted shopping becomes a larger share of how consumers research and buy.
The Signal Brief · Mariano De Vitto — Head of Marketing, Barcelona