If you rely on GA4 alone to measure the impact of AI SEO, you are navigating with a broken compass.
Don’t misunderstand me. It’s a logical launch pad. But to understand how audiences find, analyze, and ultimately choose brands, measurement must go beyond the boundaries of Google tools.
SEO is a journey, not a destination. If you’re only preparing for a specified tour, large parts of that tour disappear from view.
Sessions are the result. They can’t put together the sets of assumptions that are increasingly saved by algorithms and AI well before the visit takes place.
You can lose potential customers in the Bermuda Triangle of traditional SEO. Use the power of voice sharing to guide user intent. Target yourself by mapping your brand visibility to AI analytics.
Measuring AI visits with GA4
Links are becoming more and more common in AI systems. Traffic is increasing. GA4 makes it easy to set up a custom report to track these sessions.
Create a test with “session source / medium” as the dimension and “sessions” as the metric. Then apply this regex filter to the referrer:
.*(chatgpt|openai|claude|gemini|bard|copilot|perplexity|you.com|meta.ai|grok|huggingface|deepseek|mistral|manus|alexaplus|edgeservices|poe).*
Don’t worry if the output report is dirty. That’s normal. Most AI systems send multiple sets of component transmission information. Some don’t send anything, so the sessions appear as dark traffic.
This report is an easy first step. But don’t be fooled into thinking it can measure AI’s impact on your brand alone.
The most viewed AI results – Google’s AI overview and AI mode – cannot be seen here. They are listed in “google / organic” or “(direct) / (none),” depending on how the user accessed Google.
With these limitations, looking only at GA4 traffic from AI generators is not a complete data source to understand the reality of your target audience’s usage and impact on your brand.
Other data sources are needed.
Dive deep: LLM development in 2026: Tracking, visibility, and what’s next for AI discovery
Your customers are searching everywhere. Make sure it’s your product he appears.
The SEO toolkit you know, and the AI visibility data you need.
Start a Free Trial
Start with
Bing Webmaster Tools technically reports Copilot data. But in a very Microsoftesque way, the conversation is combined with web metrics, obscuring the conversation data and making the report ineffective in understanding the impact of productive AI.
This approach laid the groundwork for Google Search Console to do the same. AI Overview and AI Mode impressions and clicks are integrated with Search, and the Gemini app is not included at all.
What you can do is look for some conversation-style queries using the Google Search Console regex, like this:
^(who|what|whats|when|where|wheres|why|how|which|should)b|.*b(benefits of|difference between|advantages|disadvantages|examples of|meaning of|guide to|vs|versus|compare|comparison|alternative|alternatives|types of|ways to|tips|pros|cons|worth it|best|top)b.*But this is becoming less important as question fan-out becomes the norm, making artificial questions different from human questions while increasing the number of impressions.
Worse, both GSC and BWT will be increasingly myopic as websites bypass the MCP interface or are accessed directly by AI agents.
Also, other data sources are needed.
Get the newsletter search marketers rely on.
AI agent statistics by log files
Both Google and ChatGPT offer AI agents that can browse and, with permission, convert for a person.
If the AI agent uses a text-based browser, it cannot be tracked with cookie-based statistics.
When the agent switches to a virtual browser, it usually accepts cookies, 78% of the time in my testing. But this causes problems in GA4:
- Engagement metrics are not common. This is the behavior of an agent, not a human.
- Unnatural resurgence of desktop traffic. Agents use desktop browsers exclusively.
- An uptick in Chrome. Chromium active agents.
On the other hand, agency conversions are recorded, but attributed to direct traffic.
Because of this, many SEOs turn to the bot log, where the requests of the AI agent can be identified. But those requests aren’t the bulk of people sending agents to complete jobs.
When an agent renders a page in a virtual browser, it fires multiple requests for all assets. CSS. JS. Pictures. Fonts. A bloated front equals inflated claims statistics, making raw volume a null metric.
Understanding is not in the context, but in the directions.
Follow the request flow on the site to the conversion success page. If there are a lot of requests but none of them make it to the conversion path, you know the journey is broken.
See the complete picture of your search visibility.
Track, optimize, and win in Google search and AI from one platform.
Start a Free Trial
Start with
Dig deeper: How to segment traffic to LLMs in GA4
Traditional SEO reporting is not up to the task of AI tracking
To track the impact of AI SEO, you need to reevaluate your reporting.
Its benefits go beyond the boundaries of GA4, Google Search Console, and log file analysis, all of which consider that the user accessed your website directly from the AI environment. That is not necessary for the product value.
Many SEO tools now add AI tracking, or are built around it. The methodology is not exhaustive, with the results of the discussion possible, not deterministic. It is similar to using focus groups.
With unbiased sampling, unbiased information, and general testing, the resulting trends are significant, even if individual responses are not. They reveal a collection of products that the AI system associates with a specific purpose, creating a consensus view of a reliable set of assumptions.
But not all AI search analytics tools are created equal.
Make sure your tool tracks not only website citations, but also brand mentions in conversation and citations of brand assets such as social media profiles, videos, map listings, and apps.
This is less important than a website link. Seeing this shows the growth of SEO.
As an industry, we return to marketing KPIs such as share of voice. Understanding brand visibility for the right targets is ultimately what drives market share.
It is not SEO’s job to make a website better. It’s about building a well-known, highly rated, and trusted digital brand. That is the basis of appearance in every living space.
Contributing writers are invited to create content for Search Engine Land and are selected for their expertise and contribution to the search community. Our contributors work under the supervision of editorial staff and contributions are assessed for quality and relevance to our students. Search Engine Land is owned by Semrush. The contributor has not been asked to speak directly or indirectly about Semrush. The opinions they express are their own.
