There are a number of “non-test” ways to estimate FTP. Here are two threads discussing one option and the ideology behind it. This and others (WKO and more?) use certain “max efforts” to estimate FTP. These likely need to come from things other than typical workouts. Things like races, hard group rides or deliberate efforts outside of “training” mentality seem necessary for these models.
I have no real idea how accurate or useful they are, but the Intervals.icu one seems reasonably close to my Ramp test results from 2020.