Skip to content

Visibility Measurement

Measure whether AI answers name you

Registered questions are put to five engines and every answer is kept in full. Answers that name you, answers that do not, and answers that could not be measured are counted apart, never melted into one percentage.

Check technical readinessGet a quote

Free technical readiness check · no sign-up · no card

The answer and the engines

Example
Question

Which accounting software would you recommend for corporate teams?

Sample Brand A
  • ChatGPT · named
  • Gemini · measured, not named
  • Perplexity · named
  • Claude · could not measure
  • Google AI Overview · cited as a source
Example view · drawn with a sample brand and sample data, not a real measurement.

Answers are collected on 5 engines, each one reported on its own:

  • ChatGPT
  • Gemini
  • Perplexity
  • Claude
  • Google AI Overview

One question, 5 engines

Example
  • ChatGPTmeasured, not named
  • Gemininamed
  • Perplexitynamed
  • Claudecould not measure
  • Google AI Overviewcited as a source

Each engine's result stands on its own; none is blended into one score.

Example view · drawn with a sample brand and sample data, not a real measurement.

01

Each engine measured and reported on its own

ChatGPT, Gemini, Perplexity, Claude and Google AI Overview do not give the same answer to the same question. Each engine's result is shown with its own answer count; we do not blend the engines into one score.

  • If the engines did not answer the same questions, the report says so and does not rank them against each other.
  • Answers to the same question are read side by side, engine by engine; none stands in for another.

Three reasons for low visibility

Example
  • Not there at allNeither your name nor your site is in the answer.
  • Read, not namedYour site was cited, your name is not in the sentence.
  • Another companySomeone with the same name is being described.
Example view · drawn with a sample brand and sample data, not a real measurement.

02

Low visibility is three different problems

Your name may not appear at all, your site may be cited while your name goes unsaid, or another company with your name may be the one described. Each needs a different fix, so answers carry different states.

  • An answer describing a namesake company is set aside as ambiguous and is not counted as a mention.
  • For an answer that does not name you, its source list shows whether your site was cited.

The answer itself

Example
Perplexitynamed

“Options that stand out for corporate teams include Sample Brand A and Sample Brand B. Sample Brand A is mentioned especially for multi-branch support…”

Sources this answer cited

  • samplebrand-a.comyour site
  • sample-guide.comguides and comparisons
  • sample-forum.netforum
Example view · drawn with a sample brand and sample data, not a real measurement.

03

The full text of every answer, and the sources it leaned on

Every figure in the report opens down to the answers: which question, which engine, the answer's own words and the addresses it cited. If the source list could not be collected, the report says so; an empty list does not mean 'no sources'.

  • Answers filter by engine, question type, audience and mention state; the filter is shared as a link.
  • Quotes are the answer's own words, never summarised.

Named alongside you

Example
  • Sample Brand Bin price questionsnamed together
  • Sample Brand Cin trust questionsnamed instead of you
  • Sample Brand Din comparison questionsnamed together

Being named together is an observation; it does not say why the engine chose.

Example view · drawn with a sample brand and sample data, not a real measurement.

04

Who AI names alongside you

Brands, institutions and products named in the same answer are recorded as co-mentions. Question by question, you see where another brand was named instead of you.

  • A co-mention is an observation; it does not say why the engine chose a brand.
  • You set the competitor list in the brand settings.

How you are spoken of

Example
  • positive
  • neutral
  • negative

Over answers that named you only; an answer that does not name you is not counted as neutral.

Example view · drawn with a sample brand and sample data, not a real measurement.

05

How you are spoken of

In answers verified to name you, positive, neutral and negative language is told apart. An answer that does not name you is not counted as neutral; tone is read only over the answers that name you.

  • The tone distribution is not a reputation measure and is not presented as one.
  • If no answer names you, tone is not calculated, and the report says so.

Report summary

Example

Summary

Sample Brand A is named behind its competitors in comparison questions

  • Visibility in buyer questionsmeasured
  • Mentions in price questionsmeasured · no mention
  • Newly added audiencenot measured yet
Example view · drawn with a sample brand and sample data, not a real measurement.

06

What could not be measured is not a zero

When an engine does not answer or a call breaks off, that measurement is not counted as 'you were not named'. It is marked as not measured and enters no rate; the report's scope section states how many measurements were made and how many could not be.

  • Measured, measured but not named, not yet measured: each state has its own mark.
  • The report's first screen states the scope and limits of the measurement.

How it works

From question to report in three steps

  1. We set the questions together

    The questions your buyers would ask go into a registered question set; the report shows why each one was asked.

  2. We ask the engines

    Every question goes to every engine; answers are kept in full with their sources. Failed calls are recorded and not counted.

  3. We publish with the evidence

    Every claim in the report opens down to its answers. The next measurement uses the same questions.

Method promise

What we measure and what we do not guess stay apart

Visibility measurement counts the answers given to registered questions. It does not measure the engines' inner preferences, traffic from AI, or the world outside the questions; the report states these limits on its first screen.

Read the method

What we measure

  • In which question, on which engine, you were named
  • The full answer text and the sources it cited
  • The brands named alongside you
  • Measurements that failed, counted apart

What we do not measure

  • Why an engine chose a brand
  • Traffic or sales that come through AI
  • A single score blending the engines

Other modules (9)

One measurement, answers to other questions

Frequently asked questions

Which engines are measured?

ChatGPT, Gemini, Perplexity, Claude and Google AI Overview. Each one's answers are kept and reported on their own; the engines are not blended into one score.

Where do the questions come from?

From a registered question set made of the questions buyers in your category would ask. The questions stay fixed through the measurement, and the report's answer list shows every one of them.

Doesn't naming my brand in the question inflate the result?

It would, so the two are kept apart. Visibility is calculated from buyer questions that do not name you. Control questions that do name you only show whether the engine recognises you, and they stay out of the visibility figures.

What happens if an engine does not answer?

That measurement is marked 'not measured' and enters no rate. It is never counted as 'you were not named'; the report's scope section lists the measurements made and those that could not be.

How often is it measured?

The registered measurement repeats with the same question set, and two measurements are compared only if they used the same questions. For the scan-based analysis, a weekly or monthly refresh is set in the brand settings.

First, make sure your site can be read

The free check measures whether AI crawlers can reach your site. To scope a visibility measurement together, ask for a quote.

Check technical readinessGet a quote

Free technical readiness check · no sign-up · no card