How to Tell Whether AI Search Is Actually Sending You Customers
If you want to know whether AI search is actually sending you customers, the first problem is that your own analytics are probably hiding the answer. When someone asks ChatGPT or Perplexity for a recommendation and then clicks through to your site, that visit often arrives with no referrer attached, so Google Analytics 4 quietly files it under Direct traffic instead of naming the source. In one 2026 analysis of more than twenty thousand AI-driven visits, 70.6 percent landed with no referrer and defaulted to Direct (Loamly, 2026)<a href="#ref-1" id="cite-1" class="cite"><sup>[1]</sup></a>. That is the gap this guide closes. To measure AI search traffic you do not need a new analytics platform. You need to teach the one you already have to recognize the assistants, then read the numbers honestly. This matters more for a local service business than most owners realize, because a family that just moved to town increasingly asks an assistant who to call before they ever open a map. If those visits are landing in your Direct bucket, you are drawing the wrong conclusion, deciding AI sends you nobody when it may already be sending some of your best-qualified leads. The fix is a one-time setup in GA4, a clear-eyed look at conversion, and the discipline to keep watching a channel that is still young and still growing.
To measure AI search traffic, build a custom channel group in Google Analytics 4 that captures referrals from the major assistants, then compare their conversion rates against organic search. Because most AI visits arrive with no referrer and hide inside Direct, this reporting work, part of any serious SEO reporting strategy, is what turns an invisible channel into a number you can actually act on.
Why GA4 Files Most of Your AI Traffic Under Direct
Most AI referrals reach your site stripped of the data that says where they came from. Assistant apps often pass no referrer, and people frequently copy a link out of a chat rather than click it. Google Analytics 4 has nothing to attribute the visit to, so it defaults to Direct and your AI channel disappears into it.
Start with the mechanism, because it explains why the channel looks empty when it is not. A referrer is the piece of data a browser normally sends to tell a site which page a visitor came from. Many AI assistants, especially inside their mobile apps, do not send one, and a large share of users copy the link out of a chat and paste it into a browser, which sends no referrer either. In that same 2026 dataset, 70.6 percent of more than twenty thousand AI-driven visits arrived with no referrer at all (Loamly, 2026)[1]. Google Analytics 4 cannot label a source it was never handed, so it files the session under Direct alongside your bookmark and address-bar traffic.
For a local business the practical damage is a wrong conclusion. If you glance at GA4, see almost nothing under a referral or organic label tied to AI, and decide that assistants are not sending you customers, you are reading an artifact of the tooling rather than the truth. The visits may be there, sitting in a Direct bucket that has quietly swelled over the past year. That is why looking past the default report matters, and why it belongs in the same broader shift toward learning to measure SEO beyond organic traffic rather than trusting a single top-line number.
It is worth knowing that GA4 added a native AI Assistants channel in 2026 that auto-tags some recognized assistant referrers. It helps, but it does not solve the problem, because it can only fire on a visit that actually carries a referrer it recognizes. Every no-referrer session, which is the majority, still lands in Direct, and some platforms are not covered at all. So the native channel is a useful floor, not a complete answer, and building your own channel group on top of it is still the honest way to see the whole picture.
How to Build an AI Referral Channel Group in GA4
The core fix is a custom channel group that groups every recognizable assistant referral into one AI Search channel. It takes a few minutes in the GA4 admin, it applies to your reports going forward, and it turns a scattered handful of referral sources into a single line you can track, segment, and compare over time.
In GA4, open Admin, then Channel groups, and create a custom channel group. Add a new channel, name it something like AI Search, and set its condition so that the session source matches the assistant domains. A source-matches-regex rule covering chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com, and claude.ai will capture the referrals that do arrive with a referrer, and you can extend that list as new assistants appear. Order the channel above Organic and Direct so a qualifying visit is claimed by AI Search first. From then on, your reports carry an AI Search line you can read at a glance.
Be honest about what this does and does not recover. The channel group correctly labels the referred visits that carry a source, which is real and immediately useful, but it cannot retroactively rescue the no-referrer sessions that already fell into Direct. Two habits help close the rest of the gap. First, watch your Direct channel for the unexplained growth that often shadows a rising AI presence, and treat a sustained Direct climb as a signal to dig, not as noise. Second, where you have access to raw server logs, the assistant crawlers and some referrers show up there even when the browser sends nothing to GA4, so logs are a useful cross-check against the analytics view.
Finally, add UTM parameters anywhere you actually control a link an assistant might surface, such as a profile, a directory listing, or a syndicated post, so those clicks arrive already tagged. You will never tag the organic mentions an assistant generates on its own, but tagging what you can removes guesswork from that slice. Treat this setup as part of your standing measurement stack rather than a one-off, because the assistant landscape keeps shifting and the channel group is only as current as the domains you keep in it.
Which AI Assistants Actually Send You Visitors
Once the channel is live, the traffic is lopsided. One assistant dominates the referrals, a couple send a meaningful trickle, and the rest are rounding errors for now. Knowing the split tells you where to focus, because being recommended by the platform that sends most of the visits is worth more than spreading effort evenly across all of them.
The concentration is stark. ChatGPT alone accounts for roughly 87 percent of AI referral visits across measured sites (Lantern, 2026)[2]. Perplexity and Gemini send smaller but real streams, and the practical read for a local operator is simple: if an assistant is going to name your business to a searcher this year, it is most often ChatGPT doing the naming. That does not mean you ignore the others, but it does mean your first question when a lead says an assistant recommended you is usually which one, and the answer is usually the same.
The reason this is worth measuring at all is that the behavior is now mainstream, not fringe. In the past year, 45 percent of consumers used an AI tool to find a local business, up from 6 percent a year earlier (BrightLocal, 2026)[3]. For a service business, that increasingly includes the exact moment you most want to be present, when a new resident asks for a plumber, a dentist, or a med spa before they have a single local contact. The channel group is how you find out whether you are already the answer, and the platform split is how you decide where to earn more of those recommendations.
What earns the mention is the same reputation work that wins traditional local search, which is the reassuring part. Assistants lean on well-attested, third-party corroborated businesses, so the accurate profile and steady reviews you already build do double duty. If you want the deeper version of how to become the business an assistant names, our guide to how to show up in ChatGPT and AI search covers it, but the measurement point stands on its own: you cannot improve what you refuse to look at.
What AI Search Traffic Is Worth Once You Can Measure It
The reason to bother measuring AI traffic is that, visit for visit, it tends to convert well above organic search. The searcher did their comparing inside the chat before they ever clicked, so they arrive closer to a decision. The exact multiple varies widely by study and industry, but the direction is consistent across the data.
Look at conversion, not just counts, because that is where AI traffic earns its keep. In one 2026 benchmark, ChatGPT referrals converted at 15.9 percent and Perplexity referrals at 10.5 percent per session, compared with roughly 2 to 3 percent for organic search (AirOps, 2026)[4]. The intuition behind the gap is that an assistant hands you a visitor who has already narrowed the field, so a larger share of them are ready to act rather than still browsing. That is exactly the kind of high-intent visitor a service business wants, which is why counting these visits without also reading their conversion rate misses the whole point.
Treat the size of the advantage as a range, not a single number, because the honest data does. One widely cited benchmark placed AI referral traffic at about 4.4 times more valuable than organic on average (Semrush, 2025)[5], while Ahrefs, measuring its own site, found AI search visitors converting roughly 23 times better, where just 0.5 percent of traffic drove 12.1 percent of signups (Ahrefs, 2026)[6]. Those are different businesses with different products and definitions, so the gap between them is expected. The takeaway is not a magic multiplier, it is that across independent measurements AI traffic converts several times better than organic, and a local business should expect a real premium even if it will not match a software company's exact figure.
One caution from doing this in practice: early AI numbers are small numbers, so read them as a trend, not a verdict. A handful of AI-referred sessions in a month can post a dramatic conversion rate that swings wildly the next month simply because the sample is tiny. The discipline is to watch the channel accumulate over a few months before you draw conclusions, the same way you would with any new source, and to pair the conversion view with the reputation signals that feed it, since the reviews that shape AI recommendations are also what lifts the volume you are measuring.
How Measurement Points You Toward More AI Visibility
Measurement is not the finish line, it is the instrument that tells you what to do next. Once AI traffic is a visible, trackable channel with a known conversion rate, you can decide whether it deserves more of your attention, and you can tell whether the work you do to earn assistant recommendations is actually paying off.
The first thing a clean AI Search channel buys you is a defensible decision. Instead of arguing about whether AI search matters in the abstract, you can look at how many sessions it sent, what they converted at, and how that compares with the channels you already fund. For most local service businesses the volume is still modest, so the honest framing is that AI is a fast-growing, high-quality channel worth positioning for early, not a reason to abandon what already works. Measurement is what keeps that judgment grounded in your own numbers rather than someone else's headline.
The second thing it buys you is a feedback loop. The moves that earn assistant recommendations, a consistent and accurate profile, steady recent reviews, and clear content that answers the questions buyers ask, are the same fundamentals that win local search generally. When you invest in those and your AI Search channel climbs a quarter later, you have evidence the work landed. That is the practical link between being findable and being chosen, and it is why the effort to get named by AI search should be measured, not taken on faith.
None of this requires a bigger budget so much as a habit. Set up the channel group once, check it alongside your other sources on whatever cadence you already report on, and let a few months of data accumulate before you act on it. The businesses that will benefit most from AI search over the next few years are the ones watching the channel while it is still small, because they will see the shift coming in their own numbers rather than reading about it after the fact. You cannot manage a channel you refuse to look at, and now you have no reason not to look.
Turn on what makes AI recommend you.
AI recommends the businesses it can read, trust and quote. Flip on the four signals we engineer, and watch your visibility climb and the answer rewrite itself.
Illustrative · the four signals are the real system we build
This article, answered.
The questions readers ask about this topic, answered the way an answer engine would. No forms, no sales pitch.
Pick a question on the left and you’ll get the direct answer, the way an answer engine would give it.
PUBLISHED July 24, 2026 · WRITTEN BY JAMIE KLONCZ, FOUNDER · SEO ELITE AGENCY, NAPLES FL
Enter a path and click verify.