One month
The prompt set stays frozen so the months compare.
Measure
The frozen set runs again with the same number of runs. Questions naming you are reported separately.
OutputMonthly reading
Interpret
A person reads the answers. Movement beyond the spread? Your action, a competitor, an engine change?
OutputFindings with evidence labels
Implement
The next items on the list, in your system. Every change dated.
OutputChange log
Report
A report for the board and a review meeting. If nothing was measurable, it says so.
OutputMonthly report
Four evidence labels
- Observed
The engine returned this result on this date in this number of runs.
- Correlated
Two figures moved together. A hint, not proof.
- Attributed
A rule in your analytics assigned a visit to an AI source. As good as the rule.
- Causal
The measurement design supports the claim that an action caused the change. Rare.
The frame
- Prerequisite
A baseline reading. Without it there is nothing to compare a month against.
- Rhythm
Monthly: reading, implementation, report, review.
- Working with
Your marketing and development teams. One fixed contact, no queue.
- Term
Set in the proposal. Engines react at different speeds, some in days, some in months.
- What you keep
All reports, all raw observations, the prompt set, every documented change.
What to expect on timing
Perplexity fetches sources live and can reflect a change in days. Google AI Overviews follow your Google visibility. ChatGPT has an index and a cache and takes longest. Experience, not promises.
That is why we measure trends, not points. From 4 to 5 to 9 percent over three months is a trend. From 4 to 12 to 3 is spread, and we call it that.
Support includes monitoring of false statements. Details under False AI statements.
Who it is for
Worth it if
- the baseline showed gaps that take months to close
- nobody in-house should own the topic permanently
- competitors in your category are visibly active
- false statements about you have already appeared
Not needed if
- you are already stably represented in your category
- the action list is short and your developers can do it
- you want a tool to check things yourself
Which rhythm fits you?
With a baseline we know your situation. Without one, we start there.
A reply within one working day, from the person who measures.