~70% of AI-Driven Traffic Goes Untracked. Here's Where Your Revenue Is Actually Going in 2026.

AI engines cite your brand. Publishers route the click through Skimlinks. Amazon collects the sale. Here's why roughly 70% of AI-driven traffic disappears from attribution — and what it takes to see it.

Bharad Ramesh

CEO & Founder, Aqxle · June 2026 · 8 min read

Your brand gets cited in a ChatGPT answer. Perplexity recommends your product. You see the AI mention, maybe even celebrate it. Then you check your analytics — and nothing moves.

That gap is not a tracking glitch. It's a structural problem. Multiple 2025–2026 industry analyses now estimate that roughly 70% of AI-driven traffic is misclassified or invisible to standard analytics¹² — most of it landing in the "Direct" bucket of GA4 with no source attribution. For enterprise brands with meaningful publisher-driven demand, the resulting revenue leakage routinely reaches seven figures annually.

The 70% Problem

Most AI-driven demand never appears in your attribution stack the way it should. The traffic exists. The intent is real. Shoppers are asking ChatGPT, Perplexity, and Gemini what to buy, and those engines are naming brands. But the path from AI answer to purchase runs through a chain of intermediaries your current tools were never built to follow — and by the time a click reaches a merchant, the AI origin is gone.

Here's what that chain actually looks like:

1. A shopper asks Perplexity for the best wireless headphones under $200.
2. Perplexity cites a review from a major tech publisher like PCMag.
3. The shopper clicks the publisher's link.
4. That link is wrapped in a Skimlinks or Howl redirect — an affiliate monetization layer the publisher uses automatically.
5. The redirect resolves to a merchant. Most often Amazon. Rarely your DTC site.

Your brand got the citation. Amazon got the sale. Your dashboard recorded nothing.

Why Your Current Tools Stop Too Early

The tools most CMOs and Heads of DTC rely on today were built for a different problem.

SEMrush, Ahrefs, and Similarweb are excellent at tracking organic search visibility. They can tell you where your brand ranks, how much traffic a publisher receives, and which keywords drive clicks. What they cannot do is follow a Skimlinks redirect chain to its final destination.

AI visibility platforms like Profound and Gumshoe AI have moved into the citation-tracking space. They can tell you how often your brand appears in ChatGPT or Gemini answers, which is genuinely useful. But they stop at the mention layer. Neither platform currently markets the capability that actually matters for revenue: when AI recommends your brand and a shopper clicks through, where does that shopper land?

Affiliate platforms like Skimlinks sit on the other side of the equation. They see the redirect. They know where the click went. But they're publisher-facing tools, not brand-facing ones. They have no reason to tell you that 81.8% of a top-cited tech publisher's outbound affiliate links resolve to Amazon rather than your site.³

Your dashboards stop exactly where the money moves.

The Redirect Chain Problem, Explained

Publishers monetize their content through affiliate networks. When a tech review site recommends a product, the outbound link almost never goes directly to the merchant. It passes through one or more redirect layers: Skimlinks, Howl, Impact, Rakuten, or a raw redirector the publisher controls.

Each layer obscures the final destination from standard analytics. UTM parameters and click IDs frequently do not survive these redirect chains.⁴ Your attribution model sees a click leave the publisher and then loses the thread entirely. Free-tier ChatGPT users compound the problem — they send no referrer header at all, so the click lands in GA4 as "direct," the same bucket as someone typing your URL by hand.⁵

The result: a shopper who read a Perplexity-cited review, clicked a link, and bought your product on Amazon is invisible to your DTC analytics. That sale looks like organic Amazon revenue. You have no idea AI drove it. And you have no way to know whether the next hundred shoppers who followed the same path bought your product or a competitor's.

This is marketplace capture in its most insidious form. Amazon isn't stealing your demand through advertising or price competition. It's capturing demand you earned through AI citations, routed through publisher affiliate stacks you didn't set up and can't currently see.

What "AI Search Traffic Attribution" Actually Requires

Solving this problem requires connecting three data layers that currently live in separate tools with no shared view:

No single tool in the market currently connects all three layers. Brands running Profound alongside SEMrush alongside their affiliate platform data are still working with three disconnected views and no unified answer.

That's the gap Aqxle was built to fill. The Commerce Graph aggregates AI visibility data from existing tools, joins it to publisher outbound link data, and resolves every affiliate redirect to its true final merchant destination. The output is a single view of the entire path from AI answer to sale, delivered within five business days of onboarding.

If you want to see where your brand's AI-cited traffic is actually going, a free domain audit will show you your specific leakage exposure before you commit to anything.

What Brands Find When They Look

The numbers are not comfortable. When Aqxle mapped 1.85 million outbound affiliate links from PCMag — a publisher consistently ranked in the top 10 AI citation sources for technology queries⁶ — 90.6% routed traffic away from brand DTC sites entirely. 81.8% resolved to Amazon.³

If your brand appears in that publisher's reviews and you sell on both Amazon and your own DTC site, the math is stark: roughly four out of five AI-influenced clicks that publisher generates go to Amazon, not to you.

The revenue implication depends on your margin structure. For most CPG, electronics, and beauty brands, the difference between a DTC sale and an Amazon sale is meaningful: higher margin, better customer data, more control over the post-purchase relationship. Losing that sale to Amazon isn't just a revenue event. It's a customer acquisition event that benefits Amazon's flywheel, not yours.

And this is before accounting for clicks that route to a direct competitor. In categories where multiple brands compete for the same AI citations, redirect chains don't always stay within your brand's orbit. A shopper who clicked your name in a Perplexity answer can end up on a rival's product page because the publisher's affiliate stack optimizes for commission rate, not brand loyalty.

The Measurement Gap Is a Revenue Problem

There's a tendency to frame AI search visibility as a marketing problem. Get more citations. Optimize your content for AI engines. Improve your answer engine optimization score.

That framing is incomplete. Visibility without attribution is just vanity. The question isn't whether AI is mentioning your brand. The question is whether those mentions are generating revenue for you or for someone else.

Roughly 70% of AI-driven traffic goes untracked.¹² The portion that does get tracked is heavily biased toward marketplaces — in the PCMag audit, 81.8% to Amazon alone.³ Until you can see the full path, you're making content and distribution decisions based on a fraction of what's actually happening.

The measurement infrastructure for AI search traffic attribution exists now. The brands building it in 2026 will have a data advantage that compounds. The ones waiting for their existing dashboards to catch up will keep celebrating citations while their competitors collect the sales.

Frequently Asked Questions