Use NVDA to help train Speech Recognition?

I am trying to help someone start using a Windows computer with speech input and audio output.  They has lost ability to read owing to macular degeneration. The main applications are MS Word, Outlook, Edge and Zoom.

One snag would seem to be training the speech recognition program.  I remember using Dragon about 10 years ago and there was a lengthy training that required reading text to the microphone.  Windows Speech Recognition (WSR) also has training that requires reading text.  WSR depends on the user correcting mistakes properly. I hope the read back function of NVDA will help for that.

Has anyone used NVDA to read out the training text so the user can repeat it verbally to the speech recognition program (WSR)?  I expect this would only work with ear phones on.

I would appreciate information about forums that discuss integration of WSR and NVDA to use Windows Office products. I am aware of Knowbrainer.

Thanks.  This is a great forum!