How we measure AI visibility

Why we publish this

Anyone can screenshot one AI answer. That isn't a measurement.

AI answers change from one run to the next. One screenshot can make a business look invisible, or make it look like it's winning. So we'd rather show the method. If you hire us, this is what you're paying for. If you don't, you can run it yourself.

The four engines

We ask the same questions on:

• Google AI Mode
• ChatGPT
• Gemini
• Perplexity

Each one builds its answer differently. A business can be named by one and missing from the other three. We don't report one engine as if it speaks for all of them.

The query set

Every client gets a fixed list of questions. It's written before the first run and it doesn't change after.

It has two parts.

Category queries are what a buyer asks before they know your name. Things like "best med spa in Scottsdale" or "who does design-build remodels in Phoenix." This is where the gap usually is.

Brand queries use your name. Things like "is [your business] any good" or "[your business] reviews." They show what AI says about you once someone already knows you exist.

Category queries matter more. A brand query almost always finds you and tells you very little.

Why every query runs 3 times

One run isn't reliable. In our baseline testing, Google AI Mode gave a different list for the same question an hour apart.

So we run every query 3 times on every engine. Each run starts in a fresh session, so an earlier question in the same chat can't shape the answer.

What we measure

For every query, on every engine, on every run, we record:

• whether your business was named
• who got named instead
• the sources the engine pointed to, when it shows them
• the date and the exact wording of the question

Then we report how many of the 3 runs named you. 3 of 3 is a different position than 1 of 3. We don't round either one up.

After the fixes

The baseline shows what's keeping AI from naming you. Things like:

• a business AI can't confirm is one business
• a phone number or address that changes from site to site
• service pages that bury the answer
• no outside source that says what you say

We fix those first. If you want the reasoning behind each one, read how AI citations work.

30 days after the fixes ship, we run the same query set the same way. Nothing about the method changes between the two. You see both side by side.

What we don't promise

No one controls what an AI engine says. We don't guarantee a ranking or a citation.

What we can promise is that every number you see from us came from this method.

Common questions

How do you measure whether AI recommends a business?

We ask 4 AI engines - Google AI Mode, ChatGPT, Gemini and Perplexity - a fixed set of questions a buyer would ask. Each question runs 3 times, each time in a fresh session. We record whether the business was named and who was named instead.

Why do you run each query 3 times?

Because one run isn't reliable. AI engines can return a different list for the same question an hour apart. 3 runs show whether a mention is steady or a one-off.

What's the difference between category queries and brand queries?

A category query is what a buyer asks before they know your name, like "best med spa in Scottsdale." A brand query uses your name. Category queries matter more, because that's where a buyer decides who to call.

How do you know if the work made a difference?

30 days after the fixes ship, we run the same query set the same way as the baseline and compare the two. The questions and the method don't change between runs.

Can you guarantee AI will cite my business?

No. No one controls what an AI engine says. We measure it the same way before and after the work, so you can see what changed.

The Visibility Diagnostic runs this method on your business. It's $1,500, takes 2 weeks, and the fee credits toward a retainer.