AI visibility monitoring

One measurement is a snapshot. Monitoring is the record.

AI answers change. Monitoring reruns the same locked question set under comparable conditions so movement can be observed rather than assumed.

What monitoring tracks

Recommendation share
How often the team is recommended across the locked question set.
Citation share
How often the team or its pages are cited in the captured answers.
Competitor share
Which teams are recommended instead, and how that distribution moves.
Question-level performance
Which questions the team appears in, tracked question by question.
Source changes
New, lost or changed sources appearing around answers in the market.
Entity accuracy
Whether name, brokerage, markets and members are described correctly.
Provider and date
Recorded with every captured answer, because conditions change.
Answer volatility
How much the answers to the same question vary between runs.

How the two headline metrics are defined

MetricDefinition
Recommendation sharePrompts where the team is recommended ÷ total valid prompts in the captured set.
Citation sharePrompts where the team or its pages are cited ÷ total valid prompts in the captured set.

These are measurement metrics for a defined market, question set and provider. They are not universal rankings and they do not describe how any model ranks sources.

Invalid responses — refusals, non-answers, answers about a different market — are excluded from the denominator and reported separately, so a shrinking denominator can never quietly inflate a share.

What this cannot prove

  • Movement between runs may reflect model changes rather than anything you did.
  • Comparable conditions are maintained deliberately, but they cannot be guaranteed by us.
  • No monitored metric predicts future answers.

See who AI recommends in your market — and where you stand.

One client per market. Start by telling me where you work.