Progression levels on their own can’t be the predictor.
One could achieve those progression level gains in four workouts, maybe three and maybe even two.
There are multiple other factors that it must consider before spitting out its new number every month.
We could spend some time guessing what those other factors are, but volume is going to be in there for sure.
If you do knock out an FTP test, please let us know how the result compares to the AI
AI ftp is a black box, only TR will know what causes it to re evaluate things and update your ftp.
The best indicator of performance is still performance itself, to get a true indicator of FTP it’s best to go test it out to at least ~40m.
That’s not saying you can’t get close with other methods just that they have drawbacks.
And let’s remember that progression levels are a made up system that helps TR prescribe the “correct” workout from their library. It isn’t like there is published research that ties progression levels to FTP.