Can I see traffic from ChatGPT or Perplexity in my analytics?
Some of it. Clicks from an assistant's answer usually arrive as referral traffic from that assistant's domain, so you can build a segment for them. What you cannot see is the far larger number of times you were read, summarized or named without a click. Referral counts understate influence, and a meaningful share of assistant-driven visits land as direct traffic because the referrer gets stripped or the link is copied by hand.
What shows up, and what silently does not
Assistant referrers are identifiable domains, so grouping them is straightforward once you decide to. What leaks away is everything else: in-app browsers that drop the referrer, users who read your name in an answer and search for you separately, and users who copy a URL into a different browser.
The last two are the important ones. An assistant that convinces someone to look you up produces a branded search, not a referral. In your reports that shows up as organic or direct, credited to a channel that did not do the work.
Referrer values also shift as products change. An assistant that opened links in a browser last quarter may render a preview in-app this quarter, and your referral counts move for reasons that have nothing to do with your visibility. Note product changes alongside the data.
Build the segment before you need the history
Create a channel group for assistant referrers now, even if the volume is trivial. Analytics cannot retroactively classify traffic you never grouped, and in a year you will want the trend rather than a snapshot.
Annotate the date you created it and any change to how you classify. Half of the arguments about this channel come from someone comparing two periods that were measured differently.
Judge it on behavior, not on volume
Assistant referrals are usually a small share of sessions and a disproportionate share of qualified ones, because the assistant already did the research. Comparing raw session counts against organic will always make this channel look irrelevant.
The comparison that matters is conversion rate and revenue per session against your other channels, which requires sessions and job records in the same place. That join is the point of revenue intelligence, and it is the difference between cutting a small channel and understanding it.
In home services the phone matters more than the form, so this only works if calls are attributed to the session that produced them. Without that, an assistant-referred visitor who reads one page and calls looks like a bounce. Call data closes that gap.
Be explicit about the blind spot
No analytics package can tell you how often you appeared inside an assistant's answer without a click. That number is invisible by design, and it is probably the larger number.
That is why referral data and a question panel are complementary rather than redundant. The panel estimates presence; analytics measures the fraction of presence that turned into a visit. Reported together in one view, they are honest. Reported alone, either one misleads.
Topics: analytics · referral traffic · attribution · measurement
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.