Many website owners check the same metric every month: their Google ranking. But it is precisely this metric that is rapidly losing its significance. Now that Google is responding with AI Overviews and AI Mode, and more and more people are asking their questions directly in ChatGPT, Gemini, or Perplexity, the traditional ranking check captures only a small slice of reality. It’s time to rethink website tracking.
The Problem with Traditional SEO Tools
Tools like Semrush or Seobility aren’t worthless—but their numbers should no longer be considered objective truth. These providers do not have access to Google’s internal data. Rankings, visibility indices, and traffic forecasts are extrapolated from samples of search terms, assumed click-through rates, and, in some cases, clickstream data. These are estimates—and the tools in question describe them as such themselves.
As long as Google Search was predictable, these estimates worked quite well. But it is precisely this predictability that has now disappeared.
Why AI Search Blurs the Rankings
AI summaries now appear in nearly half of all Google search queries. This is changing the rules of the game in several ways:
- One query triggers many subqueries. AI mode breaks down a query into numerous sub-queries—known in technical jargon as “query fan-out.” Different sources can be consulted for each sub-query.
- The sources change. The same query may return different websites as sources in two separate searches.
- People click—or they don’t. Anyone who gets their answer right away in the AI summary won’t even visit the website anymore.
- Visible without positions 1 through 10. A page may be an important cited source in an AI response, even though it does not appear anywhere in the standard search results.
A fixed visibility metric from a third-party tool can no longer accurately reflect this. The key question, therefore, is no longer “Where do we rank?” but rather: Are we being found and used as a credible source—and does this result in inquiries or revenue?
Modern Tracking on Four Levels
Google Analytics alone isn’t enough for this, because it only measures what happens after someone arrives on the website. A modern setup consists of four levels that complement each other:
| Layer | Tool | What Is Measured |
|---|---|---|
| Google Visibility | Google Search Console | Impressions, Clicks, Search Queries, Landing Pages, Click-Through Rate, Branded vs. Non-Branded Searches |
| Website Success | GA4, or possibly Matomo | Inquiries, bookings, sales, phone and email clicks, visitor quality |
| Technical Health | Crawling, Server Logs, Monitoring | Indexing, Errors, Bots, Load Time, Availability |
| Market Comparison | Semrush, Seobility | Keyword trends, competitors, backlinks, technical audits – as a benchmark |
Search Console and Analytics are at the heart of this: Search Console shows what happens before visitors arrive at the website, while Analytics shows what visitors do on the website. Semrush and Seobility remain useful—but for technical audits, competitor monitoring, and long-term trends, not as key performance indicators.
New: Directly Measure Visibility in AI Search
In early June 2026, Google introduced its first dedicated report on visibility in generative AI features in Search Console. It reports impressions from AI overviews, AI mode, and generative Discover features separately—broken down by page, country, device, and time period.
There are two important limitations: For now, the report only shows impressions—not clicks, click-through rates, or search terms. And it’s being rolled out gradually and isn’t yet available for every website. Those who have access to it should nevertheless start tracking now—because an impressions value serves as the baseline against which all future metrics will be measured. Every month without recorded baseline values is a lost opportunity for comparison.
How much traffic comes from ChatGPT and similar services?
Here, it’s important to distinguish between traffic (someone actually clicks on the website) and visibility (the brand is mentioned without anyone clicking).
As of May 13, 2026, GA4 has a dedicated “AI Assistant” channel for traffic. It automatically tracks visits from AI assistants such as ChatGPT, Gemini, or Claude without requiring any configuration. In addition, OpenAI often appends the parameter utm_source=chatgpt.com to links from ChatGPT search results, which allows visits and the resulting requests to be tracked.
It’s important to take a realistic view: This channel systematically underreports traffic. Perplexity continues to be categorized under “Referrals”; clicks from AI dashboards are counted as organic search; and clicks from the mobile apps of AI services often arrive without a referrer and are categorized as “Direct.” A historical year-over-year comparison also doesn’t work yet, because the channel only starts counting from its launch date. The most reliable approach, therefore, remains to add UTM parameters to important links yourself—these are included in the link and remain intact even if the referrer is missing.
Mentions Without Clicks: Monitoring Your Own AI Visibility
There is still no comprehensive “Search Console for ChatGPT.” This means you can’t automatically see how often a brand has been mentioned in AI responses, in which questions it appeared, or which competitors were mentioned alongside it. You can bridge this gap by implementing your own regular monitoring:
- Define 30 to 100 relevant questions per customer—the kind that genuine prospects would ask.
- Run the same questions through ChatGPT, Gemini, Google AI Mode, and Perplexity every month.
- Record whether the customer is mentioned and whether a link to them is included.
- Document competitors mentioned at the same time.
- Rate the quality and accuracy of the answer.
- Check each question multiple times, because the answers vary.
An example for a travel provider might be: “Which Swiss tour operator is recommended for golf trips to Ireland?” The key metric is then no longer “Position 4,” but rather the AI citation rate —for example: “The brand was mentioned in 38 percent of relevant AI questions and linked as a source in 21 percent of them.”
The Modern Customer Report
A modern monthly report highlights five key metrics:
- qualified inquiries, bookings, or revenue;
- Google impressions and clicks from Search Console;
- Brand searches versus non-brand searches;
- Visitors and conversions from the GA4 channel “AI Assistant”;
- the AI citation rate based on a fixed set of questions.
The classic visibility index from SEO tools remains part of the picture—but now only as a secondary metric. After all, in the age of AI, what matters is no longer a website’s ranking, but whether it is found, used as a credible source, and generates real business results.
