GEO · Generative engine optimisation
A rate without a sample size is a guess with a font.
Do the assistants actually name you, and how often?
Every dashboard in this category will show you a visibility percentage. Almost none will tell you how many answers it came from, which is the only thing that decides whether the number means anything.
6assistants, minimum 7 runs before we report a rate
How this usually goes wrong.
- A confident number built on four answers
- Ask a question three times, get named three times, and a dashboard prints 100%. We print "not enough data yet" until the sample clears the floor.
- Yesterday’s answer, reported as today’s
- The sources an assistant cites change substantially from one day to the next, so a single check is a snapshot of a moving target, not a measurement.
- Recognition mistaken for recommendation
- A model can know your brand perfectly when asked directly and still never bring you up unprompted. Those are different questions and we keep them apart.
What happens after we find it.
Tracking runs against your own brand on a place from the waitlist. What you see on this site is sample data, labelled as such wherever it appears.
We report what the assistants said when we asked, with the sample size attached. We do not claim a cause for a change, and we never attribute a rate to work you did.
Questions, answered plainly.
Why do you refuse to show a rate sometimes?
Because below a certain number of runs the rate carries no information. We hold every rate to a floor of 7 runs of a prompt, and under that we say so instead of printing a number that looks precise.
Do you charge extra per assistant?
No. Every plan includes all 6. Most tools in this category sell the assistant you most want to check as an add-on.