Catalina seems interesting with regard to automation given its improved voice recognition.
If I define a custom voice command in Catalina which has the action to "Press keyboard shortcut" then it works well to activate a KM macro triggered by typing a hot key.
On the other hand, it would be helpful to be able to use arbitrary text fragments and not remember which keystrokes are taken or not. I can configure KM to be triggered by typing a text string. If I then manually type that text string the Macro works. Oddly however if I set up Catalina to "Paste Text" then the text is pasted into my current application with the voice command but in that case KM does not respond. Any ideas why?
It really does expand KM considerably - I have it set up now to go by voice among my 3 main applications and it is quite a nice trick indeed to do that.
That said, there seem to be some security issues or restrictions on some words. For example, I could not get "Airmail" to work by this mechanism but was able to get it working with a hotkey shortcut instead.
Also - while Voice Control seems to be quite powerful and the killer feature of Catalina IMHO, it also is quite CPU intensive.