🧭 Introduction — The problem is not that we have no data, but that we measure the wrong thing

Imagine a potential customer asking ChatGPT:

“Which three search marketing agencies should I contact in Budapest?”

AI mentions your business. The user reads the reasoning, then a few hours or days later directly types your company name into Google, opens your website and eventually requests a proposal.

What will GA4 see? Most likely a branded Google search, an organic visit or direct traffic. Yet the buying process was actually triggered by AI.

AI-based search is not simply a new traffic channel. A new information and decision layer has appeared before the buyer journey, and traditional click-based attribution models only see part of it.

To understand how to work with traditional search optimization data, read our guide on measuring SEO with GA4 and Search Console, where Roth Creative already covers the combined use of GA4 and Search Console, as well as the measurement problem of zero-click searches.

2. How has marketing attribution worked until now?

The traditional buyer journey was relatively simple:

The measurement system could track this relatively well:

• where the visitor came from;
• which campaign they were connected to;
• which landing page they arrived on;
• which events they completed;
• whether a purchase or proposal request happened.

Classic attribution models include:

• Last-click attribution: the last interaction receives all the credit;
• First-click attribution: the first interaction receives all the credit;
• Multi-touch attribution: multiple interactions share credit for the conversion;
• Data-driven attribution: GA4’s machine-learning-based approach.

These models share one assumption: that the important part of the decision process consists of measurable digital touchpoints. AI-based search, however, often performs a significant part of the research outside the website.

That is why the right question is no longer necessarily: “Which click brought the conversion?” but rather: “Which information touchpoints influenced the decision?”

To understand how search marketing is changing in the age of generative search, read our article on search engine optimization in 2026.

3. AI search separates visibility from traffic

In traditional SEO, the buyer journey looked like this:

In AI search, however:

There is not necessarily a measurable click between the two. Three typical cases:

1. Source citation: AI cites or references the website, but the user does not click;
2. Brand recommendation: AI mentions the brand while using another website as its source;
3. Information influence: the user learns the brand name from AI, then later searches for it directly on Google.

This means traffic and visibility become two separate measurement objects. In 2026, Google has already started measuring how often a website’s URLs appear in generative search features — such as AI Overviews and AI Mode — through a separate Search Console report.

To understand how search strategy changes in the AI era, read our analysis: SEO, GEO and AEO: how is search strategy changing?

4. Why GA4 alone does not solve the attribution problem

Important: we should not claim that GA4 “cannot measure AI traffic,” because in 2026 that is no longer accurate.

Google Analytics can now classify visits from certain known AI assistants — such as ChatGPT, Gemini or Claude — into a separate AI Assistant channel.

Example:

For traditional measurement, Google may be the visible source. In business reality, however, AI was the touchpoint that activated demand.

To understand how to build a transparent measurement system, read our article on the AI SEO process and reporting.

5. The new model: do not use one attribution layer, use four measurement layers

We recommend a four-layer AI search attribution model:

6. The new KPIs of “clickless attribution”

Here are the concrete metrics worth adding:

To understand why unique data and expert insights matter in AI search, read our article: why AI search rewards unique data and expert insight.

7. What should a practical AI attribution measurement system look like?

Use a simple, implementable system:

First data source: Search Console

Analyze:

• generative AI appearances, where available;
• traditional impressions;
• branded searches;
• pages;
• countries;
• trends.

Google started rolling out separate generative AI Search Console reports in a limited way from June 2026.

Second data source: GA4

Separate:

• AI Assistant;
• Organic Search;
• Direct;
• Referral;
• conversion paths.

Third data source: AI visibility measurement

Regularly run a business-relevant prompt set across multiple AI platforms.

Fourth data source: CRM

Add a field: “Where did you first encounter us?” Possible answers: Google / ChatGPT or another AI / recommendation / LinkedIn / media / other.

Fifth layer: business data

Connect:

8. Why attribution is increasingly becoming probabilistic inference

For a long time, attribution models created the illusion that every conversion could be assigned to an exact source. AI-based search shows that in many cases this is not possible.

A decision can be shaped by several factors at once:

• Google;
• ChatGPT;
• a professional article;
• Reddit;
• LinkedIn;
• a colleague’s recommendation;
• previous brand awareness.

That is why the new attribution is less like accounting and increasingly like causal and probabilistic modeling.

To understand how this connects to organic traffic, read our article on Google AI Overviews and organic traffic.

9. Summary — Invisible influence must also become measurable

The biggest measurement mistake in AI search would be to treat only what can be clicked as valuable.

The next-generation measurement system must monitor all of the following at the same time:

Companies do not necessarily need to replace GA4. They need to supplement it. GA4 and Search Console remain important parts of the measurement infrastructure, but in the AI era they must be complemented by generative visibility data, CRM data and business outcome data.

Would you like to know how visible your brand is in generative search engines — and how that connects to actual business results? Explore Roth Creative’s search marketing and AI visibility solutions.

🔗 Related reading