AI answer measurement

What AI assistants answer your customers when they ask about your category

We put those questions to ChatGPT and Gemini ourselves — repeatedly, on the free and paid tiers separately — and show you who gets named, who does not, and which pages decide it.

Request a report See a sample report

No tag on your site. No access to your analytics. Nothing to install.

Every pass, every question Sample

Best consultant in India for Australia

Best consultant in India for USA

Best for bachelors in Canada

Best for an MBA in Canada

Free tiers, three passes each

66% Share of answer across all nine passes — a real cycle, not one flattering run.

133answers in one category cycle

3×repeats per question, per assistant

362live searches the assistants ran

40cited pages fetched and read

A real finding

One number would have hidden this entirely

Same questions, same week, same consultancy — and two assistants that completely disagreed about whether it exists.

ChatGPT — free tier0 of 58

ChatGPT — Plus1 of 15

Gemini — free tier42 of 60

A single share-of-voice figure would have averaged this into something meaningless. The split is the finding: this consultancy is winning on Google's assistant and invisible on the one most of its customers use — which points at a specific, fixable set of pages rather than "do more marketing".

How it works

Three steps, nothing to install

We measure what is said about you in public, so there is nothing to connect and no data to hand over.

Your customers' questions

We build the question set for your category — destinations, cities, services, the constraints buyers actually type — and score you only on the markets you serve.

Asked repeatedly, on every tier

Each question goes to ChatGPT and Gemini, free and paid tiers separately, three times each. Repeats are what separate a real position from a coin flip.

The pages behind the answers

Every cited page is fetched and read for who it mentions. You get the list that leaves you out, ranked by how many answers each one shapes.

Read the full method

The report

What it tells you

Share of answerHow often you are named, weighted by how much each question matters, with the spread across repeats shown rather than hidden.
Free versus payingWhat a free ChatGPT user is told, against what a subscriber is told. They are frequently different answers.
Who is named insteadYour real competitive set, discovered from the answers rather than from a list anyone supplied.
Nearly wonQuestions where you appear in some runs and not others — the cheapest positions to consolidate.
The work listCited pages that never mention you, split into a rival's own site and independent pages you can actually get onto.

Walk through a real report

Method

What we will not do

Invent numbersNo estimated traffic, no projected revenue, no "AI search volume". Nobody publishes those, and guessing would make everything else untrustworthy.
Sell you noiseIdentical questions disagree between runs. We measure that churn and report anything smaller than it as flat, even when a rise would look better.
Count an answer the model made upReplies produced without a search are reported separately and never scored — they describe training data, not what your customer is told today.
Claim credit we cannot proveNo tool can honestly connect AI visibility to revenue yet. We show visibility and the referrals in your own analytics, and let you draw the line.

Start here

Find out what you are being told about

Send us your company and category. We will run the first report and send it whether or not you become a customer.

Request a report