New

Shopify launched Agentic Storefronts. We make AI agents recommend you - not just list you.

See the Shopify integration
Tru Commerce
PricingStart free →

SOLUTION · MEASUREMENT

GA4 says Direct. It is not Direct.

A large share of agent-referred traffic arrives with no usable referrer and gets filed as Direct, so the fastest-growing channel in your business looks like noise. Server-side attribution separates it out.

How this usually surfaces

Direct traffic is climbing and nobody can explain why. Conversion rate on that segment looks unusually strong. Someone suggests it is brand awareness or an untagged campaign. Meanwhile a real acquisition channel is growing inside the bucket you dismiss as unattributable.

What is actually happening

Assistants often strip the referrer

When a session originates in a chat interface, the referrer is frequently missing or unusable by the time it reaches your analytics. GA4 has nowhere to put it, so it defaults to Direct.

The traffic converts unusually well

Agent-referred visitors arrive pre-qualified — the assistant has already answered the comparison question. That skews the Direct segment's conversion rate upward and hides the real story inside an average.

You cannot budget for what you cannot see

Undermeasured channels get underfunded. Teams keep spending against channels they can attribute while a compounding one goes unnoticed.

Client-side tagging cannot fix it

The information is missing before the page loads. Recovering it requires server-side signals and pattern analysis rather than another UTM convention.

How we work it

Instrument

Server-side collection that captures the signals available at request time, before client-side context is lost.

Separate

Agent-referred sessions are identified and split out of the Direct bucket, by surface where the signal allows.

Reconcile

We report the gap between attributed and actual, so you know the true size of the channel.

Report

Agent-driven revenue tracked alongside your existing channels, in terms your finance team accepts.

Proof

MyMuse's real AI-driven revenue versus GA4's reported figure over a 12-month attribution window

Read the case study

What does the work

Questions

Why does AI traffic show as Direct in GA4?

Because the referrer is commonly missing or unusable when a session originates inside an assistant. With nothing to attribute, GA4 falls back to Direct — so the channel exists in your data but not as a channel you can see.

How much am I actually missing?

It varies by category and how much of your demand has moved to assistants. In one 12-month attribution window MyMuse's real AI-driven revenue was 7× what GA4 reported. We measure your specific gap rather than applying an industry number.

Do I need to change my analytics stack?

No. This runs alongside GA4 and reconciles against it. You keep your existing reporting and gain a channel view you did not have.

Is this the same as DACT?

Yes — dark agentic commerce traffic, a term coined by Scot Wingo at ReFiBuy for exactly this phenomenon. Our guide to it is in the insights library.

Other jobs

Start with the baseline. Free, two fields, back in 24 hours.

Run the free Citation Rank scan or talk to founders