Method

How we measure AI visibility.

On this page

1. Start with the decision

Before choosing a single question, we define the commercial decision being studied:

FieldExample
IndustryHealth insurance
MarketThe market being studied
BuyerA family comparing cover
TriggerA change in circumstances
Job to be doneFind suitable health cover
DecisionWhich providers to look at
BrandsThe providers in scope
AI platformsChatGPT, Gemini, Perplexity, Copilot

2. Build the question set

Questions come from how buyers actually talk, not from one researcher at a desk:

  • Sales and service conversations
  • Customers' own language
  • Search and query data
  • Community and forum discussions
  • Review by specialists in the sector

Every question is tagged by buyer, market, buying stage, job, question type, whether it names a brand, and commercial importance. Most questions are unbranded, because shortlists are formed before buyers know which names to ask about.

Study typeQuestionsUsed for
Pulse20–30Quick observations
Category study75–150Detailed diagnosis
Benchmark200+Market-wide comparison

3. Collect the answers

Every question is asked on every platform several times within a defined window. For each answer we record the date, platform and interface, the exact question, the full response, citations, the brands named, the order they appear in, and how each is described. Reports state exactly what was collected, for example: “960 responses collected between 5 and 9 October 2026.”

4. The measures we report

MeasureWhat it tells you
Answer ShareHow often you are meaningfully named across eligible questions.
Recommendation ShareHow often you are presented as an option worth considering.
Shortlist ShareHow often you make the finite list of options in recommendation answers.
First-mention shareHow often you are named first when the answer lists options. This is not a ranking: AI rarely claims to rank.
Citation ShareHow often answers draw on your own content (owned) or third-party sources about you (earned).
AccuracyEach material claim about you, assessed as accurate, partly accurate, outdated, unsupported or incorrect, with the claim itself shown.
NarrativeThe descriptions AI repeats about you, with how often each appears.

5. Trace the sources

Counting mentions is not enough; the useful question is why a brand appears. We classify every cited source:

  • Your own site
  • Government and regulators
  • Industry bodies
  • News and editorial
  • Analysts and researchers
  • Comparison and review sites
  • Communities and forums
  • Social and video
  • Academic sources

Then we look at how often each type appears, how varied and how recent the sources are, and which sources mention you and your competitors.

6. State how confident we are

This is why we write “in 18 of the 40 answers that named Brand A, it was described as innovative”, not “AI sees Brand A as innovative”.

High confidence

Seen across several questions, several runs and several platforms.

Moderate confidence

Consistent within a narrower group of questions.

Exploratory

Seen once or inconsistently. Reported as an observation, not a finding.

7. What we don't do

  • We don't publish a single visibility score out of 100.
  • We don't present explanations as facts. Interpretation is labelled as interpretation.
  • We don't claim that visibility alone caused a sale.
  • We don't invent data. Illustrative examples on this site always use fictional brands and are labelled as such.

What are the limits of this method?

AI answers vary with time, location, account settings and the interface used, and models change without notice. We control what we can, document what we cannot, and report the collection window with every result. A measurement is a snapshot of a moving system; tracking the same questions over time is what makes it useful.

Frequently asked questions

Which AI platforms do you test?

ChatGPT, Gemini, Perplexity and Copilot as standard, with others added where they matter in a particular market.

Why ask each question more than once?

Because the same question can produce different answers. Averaging several runs separates consistent patterns from one-off results.

Do you use any AI to analyse the answers?

We use software to collect and organise answers at scale. The question design, the accuracy checks and the conclusions are made by people.

Related

See your Answer Share.

Your free Answer Share report tests the questions your buyers ask and shows who AI names instead.

Get your Answer Share report