Monitoring and re-measurement
The step most vendors skip, and the only one that proves anything.
Generative answers vary between runs. That variance is exactly why a fixed method matters: without one, any movement can be presented as an effect and any lack of movement can be explained away.
Our reporting is deliberately narrow: three metrics, the raw answers behind them, and what changed since last round. Short reports are harder to hide a weak method inside.
What is actually handed over.
Everything listed here is a thing you can look at, not a status update.
Three metrics
Mention rate, share of voice, and source coverage — reported per platform.
The raw answers
Every answer kept verbatim. You can check any number against its source at any time.
Change from last round
What moved, what did not, and the most likely reason for each.
Context breakdown
Positive, neutral and negative mentions counted separately. A negative mention is not a win.
Four steps, in this order.
Order matters — each step produces what the next one needs.
Questions teams ask before they start
How often should re-measurement happen?
Monthly suits most engagements. More frequent runs get lost in the natural variance of generated answers; less frequent and you cannot correct course in time.
What if the numbers go down?
We report the drop and what most likely caused it — a model version change, a competitor placement, a source that went offline. Reporting a drop is the point of measuring; a method that can only produce increases is not measuring anything.
Can we see the raw answers ourselves?
Yes. Every answer is kept verbatim and provided with each round, so any figure in the report can be traced back to the text it came from.
What does a report look like?
Two pages: three metrics per platform, the change against the previous round, the three questions that moved most with the most likely reason for each, and the context distribution reported separately from the totals.
Is there a dashboard?
Not as part of this stage. Reporting is delivered as records you can check rather than as a live figure, because a number without the answer behind it cannot be verified.
What if a model version changes mid-round?
It is recorded and stated in that round's report. A round spanning a version change is not comparable in the usual way, and we do not attribute its movement — in either direction — to the work.
Can we run the re-measurement ourselves to check?
Yes, and the method is published in full for that purpose. Run the same frozen set under the same conditions; where your numbers and ours diverge, the raw answers from both sides show which step differs.
The other stages
Find out where your brand actually stands in AI answers.
An AI visibility diagnosis across the platforms your customers use, on your real questions. No charge for the first look.