- Visibility and clicks are diverging: according to Pew Research, users click a result in 8% of searches with an AI summary, versus 15% without one.
- Since August 31, 2026, Search Console has offered every website a dedicated report for AI Overviews and AI Mode, though it shows impressions only, with no clicks or queries.
- Five groups of metrics capture your impact in AI search: visibility in AI answers, share of mentions, query mix, brand and demand metrics, and on-site user signals.
- More measurement isn't automatically better: every additional metric has to improve a decision, or it costs more than it's worth.
For a long time, a few metrics were enough to describe SEO success: clicks, impressions and ranking positions. With Google's AI Overviews and AI Mode, that no longer holds. A page can be cited in an AI answer without anyone clicking, and a falling click-through rate no longer automatically means a page is losing relevance.
If you keep looking only at the traditional metrics, it's easy to draw the wrong conclusions about the reach and impact of your content. In this article, we show which metrics are meaningful in AI search, how to measure them, and how we turn them into a KPI framework that supports real decisions.
Visibility and clicks are diverging in AI search
Google now answers many queries directly on the results page. Many brands still appear prominently there but get fewer organic visits, because users no longer need to click.
The Pew Research Center analyzed browsing data from 900 US adults in March 2025: with an AI summary, users clicked a result in 8% of searches, compared with 15% without one. As a result, click-through rate is losing value as a standalone success metric. Today, a high number of impressions can mean that content was recognized as relevant and already used in the AI answer.
Other results-page features also deliver answers without a click, such as knowledge panels or the “People also ask” box. Evaluation is therefore shifting from raw click counts toward presence, relevance and brand perception, and ultimately to the question of which inquiries and sales they generate.
Five groups of metrics to measure your impact in AI search
Beyond clicks and rankings, metrics that capture visibility, relevance and user behavior in AI-powered results are gaining importance. We group them into five categories.
Visibility in AI Overviews and AI Mode
The most important new metric is visibility in AI answers: how often your domain appears in AI Overviews or AI Mode as a source, mention or link. For Google, Search Console now provides its own data on this. For ChatGPT, Perplexity or Gemini, you track the mention rate through regular sampling of defined prompts or with specialized monitoring tools.
Your share of mentions in AI answers
Share of voice is your brand's share of all mentions in AI answers within a topic. It shows how prominent you are compared with competitors. Regular benchmarking reveals which sources the systems trust and where you can gain ground.
The query mix
Users ask longer, more complex questions in AI search. That makes it worth tracking the query mix: the share of long-tail queries, the distribution of search intents, and how both shift over time. This analysis shows which content matters most for AI search and where new opportunities are emerging.
Brand and demand metrics as context
Because visibility no longer necessarily leads to clicks, brand and demand metrics become more important: search volume for your brand name, direct traffic and returning users. They show whether visibility in AI search builds trust and demand over time.
Quality and on-site user signals
What happens on the landing page is still decisive. Conversion rates, completed tasks and drop-offs show whether visitors find what they're looking for. These metrics tie visibility to what matters for your business: inquiries, sales and their contribution to profit.
Search Console now measures AI visibility directly
Google has partly closed the gap in its own reporting. Since August 31, 2026, every website has had access to a dedicated Search Console performance report for generative AI features. It shows impressions in AI Overviews and AI Mode, broken down by page, country, device and date.
Its limitations matter when you analyze the data:
- No clicks and no queries: the report shows impressions only.
- AI Overviews and AI Mode combined: Google doesn't describe any way to separate the two formats.
- No double counting: the data comes from the “Web” search type in the standard performance report. Add both reports together and you count everything twice.
- One impression per answer: if your website appears several times in the same AI answer, the chart counts it as one impression.
In its July 2026 guidance on AI search, Google explicitly recommends using Search Console to monitor how content performs in these features. According to Google, optimizing for AI search is still SEO. So the foundation of measurement stays the same; it simply gains an AI view.
A few well-chosen tools are enough for reliable measurement
Search Console and Google Analytics are the foundation. For everything they don't cover, add these methods:
- AI answer monitoring tools: they track how often your domain is cited in AI Overviews and in ChatGPT or Perplexity answers, across languages and countries. Common SEO tools such as Sistrix, SEOmonitor and Semrush now include these features too.
- Results page sampling: regular checks of defined search terms document how AI answers and mentions change.
- Visitors from AI sources: in Google Analytics, you can analyze visits from chatgpt.com, perplexity.ai or gemini.google.com as a separate segment. Clicks from AI Overviews and AI Mode are still counted as organic search there.
- Log files and APIs: server logs show which AI crawlers access your pages. APIs let you feed the data automatically into existing dashboards.
In the dashboard, the AI metrics then sit alongside impressions, click-through rate and conversions. Because AI answers are dynamic and often personalized, we regularly review samples ourselves, compare sources and document discrepancies. That's the only way to keep the data comparable over time.
Every additional tool costs time and money. Measuring with Search Console and Google Analytics is less granular, but usually already in place. So we decide up front which metrics will actually improve a decision, and measure only those.
Treat every AI metric with healthy skepticism
Valuable as the new metrics are, they have limits. How often AI answers appear depends heavily on industry, language, region and search intent. If you compare your visibility across different keyword groups, you have to account for these differences.
On top of that come measurement errors: tools and samples return differing values due to dynamic results, personalization and location differences, and some data arrives with a delay. Only compare results that use the same measurement method and comparable data sources. Check regularly whether data collection has changed.
The biggest risk is overinterpretation. High visibility in AI answers doesn't automatically mean more demand or more profit. Without conversions, inquiries and branded searches as a cross-check, it can be a misleading signal. Some domains are also cited more often simply because they've been considered trustworthy for years, which skews competitive comparisons.
We track AI performance separately and weigh measurement costs against benefits
To make SEO manageable in AI search, you need a KPI framework that combines traditional and AI-specific metrics. We measure organic performance and AI search performance separately, and our first priority is the ratio of effort to insight.
Rather than collecting data that says little while AI systems still account for a limited share of search, we analyze how much of a website's actual traffic and inquiries comes from language models. For our clients, we maintain a clear dashboard and distill the relevant insights from it for regular review meetings. That way, you see the key developments and the decisions that follow from them.
Our GEO check also shows whether your website is visible in ChatGPT and Gemini, and which measures will increase mentions and visits from these sources.
Conclusion: Measure what drives inquiries and profit
Clicks and click-through rate still matter, but in AI search they're no longer enough. Prioritize metrics that capture visibility, relevance and quality: visibility in AI Overviews and AI Mode, share of voice, brand and demand metrics, and conversions and completed tasks.
To get started, we recommend a pilot with selected topics or page groups. It quickly shows which metrics are meaningful for your business, which tools are worth it and how they fit into your existing reporting.
