You can’t automate with the knobs only with the sliders by inserting the corresponding parameters in them..
From the manual inside system settings:
“Sounds slider mode: Step, Realtime—In SteP mode, note FX data input from Tempest’s Sound Sliders is recorded per step. This does not allow for auto- mating effects such as smooth filter sweeps. In Realtime mode, the two sound sliders affect Note FX for a sound in realtime during performance. While note FX data is still recorded in step format when sequencing, this allows you to add non-stepped, fluid, realtime control during live performance.”