AI traffic hiding in your Direct channel? The data says not yet
There’s a popular theory in SEO circles: AI platforms are quietly driving traffic that lands in your Direct channel, unattributed. The logic is plausible. Users see your brand recommended by ChatGPT or Google AI Mode, remember it, and type the URL later. No referral string. Shows up as Direct.
Kevin Indig of Growth Memo went looking for that signal across 12 Google Analytics 4 properties. He didn’t find it.
What the Data Actually Showed
Indig analyzed August through September 2026 traffic, excluding homepages, localized roots, and untracked landings. Eight of the 12 properties showed Direct traffic growth, some exceeding 100%. That looked promising until he separated engaged sessions from total sessions.
At one site, GA4 recorded 409,000 Direct sessions. Only 198 counted as engaged. The rest were bots.
Across the full January through September filtered sample, total Direct sessions rose 140.9%. Engaged Direct sessions changed -1.0% in monthly counts, or +2.3% in average daily volume when normalized for days per month. Essentially flat. The apparent growth was bots, not humans arriving from AI answers.
Direct sessions with identifiable AI traces averaged just 0.003% of total direct traffic across the 12 sites.
The Correlation That Looked Good Until It Didn’t
Indig also ran a second test: did higher AI visibility correlate with more branded search or Direct traffic in the weeks that followed? He ran 632 tests across four brands, comparing weekly AI visibility against both metrics at delays from zero to six weeks out.
One site showed a 0.72 correlation between AI Share of Voice and Direct traffic, peaking at a three-week delay. A second showed a 0.47 correlation between ChatGPT mention rate and Direct traffic. Both looked meaningful.
He then ran a validity check by deliberately mismatching the AI series against the traffic series and giving each wrong alignment the same seven chances to find a strong correlation. Several mismatched versions produced correlations as strong or stronger than the real data. The relationships also shifted by time period: one site’s correlation concentrated in the second half of the timeline, the other’s in the first half. Neither held consistently across the full window. With only 17 and 35 usable weekly observations at the selected delays, the sample size couldn’t separate a real effect from normal weekly variation.
What to Do Instead
Indig’s conclusion: click-based attribution breaks when AI is in the funnel, and the current signal in GA4 isn’t strong enough to detect hidden AI-driven Direct traffic without self-reported data. Ramp’s growth team (per Eric Wu) confirmed this approach works: they identified 30-40% of their direct traffic as originating from answer engines by examining growth in Direct visits to long blog URLs, filtering out email campaign patterns, and matching behavioral signals like time on page and scroll depth against organic visitors, then confirming with self-reported attribution surveys.
The practical measurement stack Indig recommends combines three signals: track where your brand appears across AI platforms, ask new customers directly which AI tool they used and what they were researching, and connect those answers to qualified leads and revenue. Put exposure, customer behavior, and business outcomes on the same monthly report.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
