We measure.
Capability is moving faster than most people can track. Rankings and vendor pages answer a different question than the one a business actually has.
The question here is simpler: can AI reliably perform this real task today, under defined conditions?
We are not an AI cheerleader. We are not an AI skeptic. If a test fails, the failure stays on the page. If we have not tested something, we say so.
The long-term asset is the dataset: capabilities, repeatable tests, failures, cost, time, and the evidence behind each status. The website is the public window.