Because those engagements publish none. A number that the client it belongs to has not approved for publication does not go on this site, and an invented one is worse than an empty space. 2 of the 4 studies here carry figures and 2 describe the build only.
A measure has to move against a stated baseline, and the baseline has to be the period immediately before the work started. Comparing a good month against a bad one from 2 years ago produces a bigger number and tells you nothing about what the system does.
They are the engagements with a published page, which is a different thing from either. A study exists where a client agreed to one, so this is not a ranked list and it is not the full client list.
The components repeat and the configuration does not. An AI receptionist for an electrician and one for a healthcare provider share a call flow and share almost nothing else, because what a caller wants and what a business is allowed to say back are both different.See AI Solutions
Each published study carries its own date. We do not restate a result from an engagement that has since changed shape, which is why studies leave this page as well as join it.
The situation in the client’s own terms, the specific configuration that went in, what did not work on the first attempt, and the measure with its baseline. The card here is the summary; the study is the working.