Our method

How we measure AI visibility

We measure AI visibility as your share of voice: how often AI engines such as ChatGPT, Perplexity, Gemini and Google AI Overviews mention your brand across many runs of a fixed set of buyer questions, compared with your competitors.

We show you the raw data, and we tell you where the measurement stops.

Get your AI visibility audit

Contact Jarek

The method in five steps

A fixed prompt set
Many runs, several engines
Two independent data sources
Competitors on the same prompts
Traffic and business data next to the AI numbers

Why it matters

Why “position in ChatGPT” doesn’t exist

AI answers change every time you ask. SparkToro and Gumshoe ran 12 prompts nearly 3,000 times through ChatGPT, Claude and Google’s AI with 600 volunteers. The chance of getting the same list of brands twice was less than 1 in 100; the same order, about 1 in 1,000.

So a screenshot of ChatGPT naming you proves nothing, and neither does one that doesn’t. What holds up is a percentage across many runs of the same questions, tracked over time.

That’s what we report.

Method

The method in five steps

A fixed prompt set

We agree with you on 30 questions your buyers actually ask AI, in their words:

Category questions. “Best [service] for [type of company].”
Comparison questions. “[You] vs [competitor]”, “alternatives to [competitor]”.
Problem questions. “How do I fix [problem]?”, “how much does [service] cost?”

The set stays fixed, so month-to-month changes mean something. We review it quarterly and log every change.

Many runs, several engines

Each prompt runs repeatedly across ChatGPT, Perplexity, Gemini and Google AI Overviews, in your market.

Two independent data sources

Prompt tracking follows our prompt set: whether you’re mentioned, how early, with what sentiment, and which sources the answer cites.
A broad AI answer index gives a second, independent view: how many AI responses mention your brand and which of your pages get cited, across a much larger set of queries we didn’t write ourselves.

If the two point in different directions, we say so and look for the reason.

Competitors on the same prompts

You’re always measured against 3 named competitors on the same questions. A share of voice only means something next to theirs.

Traffic and business data next to the AI numbers

AI referral traffic in your analytics (visits from ChatGPT, Perplexity, Gemini, Copilot).
Google Search Console’s generative AI performance report, available to all sites since August 31, 2026: impressions in AI Overviews, AI Mode and Discover.
Leads and revenue, where your analytics and CRM allow it.

What you get

What you see every month

Metric	What it means	Source
Share of voice	How often you appear vs competitors, across runs	Prompt tracking
Mentions	In how many answers you’re named	Prompt tracking, AI answer index
Citations	Which of your pages AI cites	Prompt tracking, AI answer index
Cited sources	Which third-party sites AI relies on in your category	Prompt tracking
Sentiment and accuracy	How AI describes you, and whether it’s right	Prompt tracking
AI traffic	Visits from AI engines	Your analytics
AI impressions	Visibility in Google’s AI features	Google Search Console
Work done	What we delivered, what’s next	DatGrowth

Limits

Where measurement stops

We’d rather tell you now than have you find out later:

Attribution is limited

Linking AI visibility directly to revenue is still hard for everyone. In the SEOFOMO State of AI Search Optimization 2026 survey, connecting AI search to traffic, conversions and revenue was among the top challenges practitioners named. We show what can be measured and label what can’t.

Prompts are a sample

No prompt set covers every question every buyer asks. That’s why we add a broader AI answer index and review the set quarterly.

Engines change

AI platforms change their sources and behavior often. A shift in your numbers can come from the engine, not from your work. When we see that, we’ll say so.

Tools estimate

The tools we use are the best we know for this, and they’re still estimates. We use two independent sources so neither is taken on faith. When a better tool appears, we switch and tell you.

Our numbers

Our own numbers

We measure DatGrowth the same way we measure clients.

FAQ

How we measure FAQ

How do you measure AI visibility?

We track a fixed set of buyer questions across several AI engines, run each one many times, and report your share of voice: how often you’re mentioned compared with your competitors. We add citation data from two independent sources, and AI traffic from your analytics and Google Search Console.

What is AI share of voice?

AI share of voice is the percentage of AI answers, across a set of prompts and runs, in which your brand appears, compared with competitors on the same prompts. It’s the most stable way to measure visibility in answers that change every time.

Why do you use two tools?

Because each one sees part of the picture. Prompt tracking follows the questions we choose for you. A broad AI answer index shows mentions and citations across a far larger set of queries we didn’t write. When both move the same way, you can trust the trend.

Can we see the raw data?

Yes. You see the same data we see: the prompts, the answers and the sources.

Can you prove AI visibility brings revenue?

Not always, and nobody honestly can yet. We report AI referral traffic and, where your tracking allows, leads and sales from it. We’ll tell you clearly which numbers are measured and which are estimates.

How often do you measure?

Prompts are tracked continuously, and you get a full report every month.

Is your method the same for every client?

The method is the same. The prompt set, competitors, engines and markets are built for each client.