Why we measured ourselves first
I built an audit that checks whether AI answer engines can see a business, and whether the numbers those engines produce are grounded in anything.
The first thing an instrument like that ought to be pointed at is the people selling it. That includes us.
So we ran it on ourselves before we ran it on the market. Our score went down. The composite fell from 62.5 to 50.0, and every number we had produced before that correction was inflated. We published the corrected figure, retired the old one, and re-ran everything.
I am telling you that because a firm that hides its own correction has already shown you how it will handle yours.
There is a version of this business where you build the dashboard first, sell the number, and add the proof later. That version is faster and it is what the market did. The problem is that proof does not bolt on afterward. A witness that does not grade, a record that cannot be altered, a signature that software cannot forge — none of those are features you add in a release.
We built them first. It took longer. It is the only reason anything on the Findings page can be checked rather than believed.
The findings themselves are on the Findings page and the method is on the Method page. Read the method before the findings if you are the sort of person who wants to know where to attack it. That is the right order, and I would read it that way myself.