There should be a feature that allows me to use my voice to issue commands. For instance, I could say, "Help me write this email," and then provide the email's content. The system would enhance that text instead of requiring me to dictate the entire email.
This can be implemented in two ways. If I press and hold a key, my voice will be transcribed as spoken. However, if I use a toggle shortcut, my voice will serve as a command rather than dictation.
The ideal scenario would be if it is able to use screen context, similar to Alter, which uses its own Appsense technology https://alterhq.com/docs#interacting-with-your-apps