By Fraser Black – Managing Director, NineFiftyNine.net

Over the last year, one question keeps coming up in monthly reports, reviews, and analytics deep dives:

“How much traffic are we getting from AI?”

It sounds like a simple question. It isn’t.

The uncomfortable truth is this: AI has already changed how people discover brands, but our analytics stack hasn’t caught up yet. Not because teams are doing anything wrong – but because AI platforms fundamentally break the assumptions web analytics were built on.

This article explores what you can see today, what you can’t, what you can reasonably infer, and how to think clearly about AI traffic without misleading yourself or your stakeholders.

The Old Model: Clicks Told the Story

For two decades, digital measurement relied on a simple chain:
Query → Result → Click → Session → Conversion

Tools like Google Analytics (now GA4) were built to answer where users came from, what they did, and whether they converted. That model worked when search engines acted as indexes.

AI doesn’t work like that.

The New Reality: AI Answers First, Clicks Later (or Never)

AI platforms summarise, synthesise, and recommend – often removing the need to visit a website at all. When a click does happen, it’s frequently delayed, indirect, or stripped of referral data.

Most AI influence does not show up as AI traffic in GA4.

What GA4 Can Actually See Today

GA4 only shows confirmed AI traffic when clean referral data is passed. In practice, this usually means a very small number of sessions, often less than 1–2% of total traffic.

This is not the full AI impact – it’s the provable minimum.

Where AI Traffic Hides Instead

AI-influenced traffic often appears as:

  • Direct traffic
  • Brand search uplift
  • Organic traffic with compressed engagement behaviour

These patterns are not anomalies – they are AI fingerprints.

Why Engagement Metrics Are Falling (and Why That’s Not Bad)

Higher intent does not mean longer sessions.

AI pre-qualifies users. They arrive informed and decisive. GA4 still rewards time spent and page depth, but AI-era users validate quickly and act.

Engagement metrics fall even as efficiency improves.

What We Can Infer with Confidence?

  • AI is absorbing informational demand
  • Influence is indirect
  • Direct traffic now includes AI impact
  • Traditional SEO under-reports visibility
  • Waiting for perfect attribution is a mistake

The Right Mental Model: Influence, Not Clicks

AI is not another channel. It’s a decision layer.

Visibility inside AI answers, brand association, and prompt-level presence matter more than raw click counts.

The Bottom Line

AI is already shaping demand, perception, and choice – even if analytics tools can’t fully capture it yet.

The companies that win will read signals instead of chasing false precision, and optimise for being chosen, not just being clicked.

AI traffic isn’t missing. It’s just no longer behaving the way analytics taught us to expect.

If you need help understanding and tracking your AI traffic. Let’s Talk.