How do I track my AEO performance ?

What do you put in an AEO tracking dashboard?

The dashboard does not need to turn all the data into an artificial "AEO score." It is more useful to keep the metrics separate because each one describes a different event.

A simple structure can be:

Google Search Console

• What you record: impressions, pages, countries, devices, and dates from Google's generative AI features
• What it shows you: how often URLs from the site appear in Google's generative AI features

Consistent test questions

• What you record: date, system, exact question, whether web search was used, whether the business was mentioned, which sources or links appeared
• What it shows you: how the answer to the same question changes over time

Analytics

• What you record: visits from AI sources with an identifiable referrer or UTM
• What it shows you: how much traffic you can technically link to a specific source

On-site actions

• What you record: forms, calls, purchases or other recorded events for the recognizable visits
• What it shows you: what those specific visitors did within the website

These elements are not equivalent. A Google impression, a ChatGPT mention, and a visit with a recorded source are different events and should remain separate metrics.

What changes with Google Search Console?

Since June 2026, Google has provided a dedicated Generative AI performance report in Search Console for AI Overviews and AI Mode. As of August 31, 2026, Google says the report has been rolled out to all websites worldwide.

The report is built around impressions and supports analysis by:

• page,
• country,
• device,
• date.

If the report does not appear for a property, Google says the usual reason is that the site has not received enough impressions in generative AI features or has been excluded from them.

This is official Google visibility data, not an "AI ranking." Use it to understand where and how often URLs from the site appear in generative AI features, and keep it separate from clicks, conversions, and other measurements.

How do you track ChatGPT and other AI systems?

For systems without an equivalent official performance report, use a consistent set of real questions.

For example:

"I'm looking for an accounting firm in Kozani for a sole proprietorship. Which firms publicly state that they support myDATA, and what can I check before I get in touch?"

For each check keep:

• the exact same question,
• the system and, where it matters, the model or feature,
• the date,
• whether web search was used,
• whether the business appeared,
• which sources or links were used,
• whether the information was correct.

This does not create an "AEO ranking." It creates a repeatable observation record so you can compare the same question over time.

Interpretation needs care: AI answers can change even when your website has not changed. That is why you should not attribute every change solely to your own actions.

How do you track visits from AI?

Here you need to distinguish "an AI mentions me" from "I have a recorded visit from this source."

ChatGPT, for example, automatically adds utm_source=chatgpt.com to referral links from ChatGPT Search results. This lets you recognize that incoming traffic in analytics tools.

For other platforms, check how the source is actually recorded in your analytics. Do not assume that every AI-related visit will appear with a clean, stable referrer.

If the source is recognizable, you can examine:

• visits,
• landing pages,
• interaction,
• forms,
• calls,
• purchases or other events you have defined.

This shows what happened after a recorded visit. It does not prove that a specific AEO change caused the visit or the conversion.

How often should you check it?

There is no single correct frequency for all businesses.

If the data is scarce, very frequent checks may create more noise than information. If there is a large volume of data, more frequent tracking may make sense.

More important than whether the check happens every week or every month is that the following stay stable:

• the core check questions,
• the systems you track,
• the fields you record,
• the time windows you compare.

For Google Search Console you can use daily, weekly or monthly analysis, depending on the volume of impressions and the question you are trying to answer.

For manual checks on other AI systems, a consistent periodic check is more useful than continuous random testing.

The broader framework of the metrics is in how you measure AEO success and in how you measure AEO ROI.

Do you need a specialized tool?

Not necessarily.

For a small set of questions the following may be enough:

• Search Console,
• Google Analytics or another analytics tool,
• a simple logging sheet for the recurring AI questions.

Specialised platforms can reduce manual work as the number of questions, systems, or brands grows. Their own scores and estimates should not be presented as an official "internal ranking" from Google, ChatGPT, or another AI system.

If you want us to examine which data makes sense to track in your own case, we start from the real questions of the customers, the available official data and what can be safely attributed in analytics.

AEO tracking can make changes more systematically measurable. It does not guarantee appearance, mention, referral, recommendation, ranking, visits or a business outcome in any AI system.