Text-to-Speech - help fixing mispronunciations

It would be great if we could record our voice to help fix and identify mispronunciation on the text-to-speech feature. I’ve had “resume” mispronounced, because I needed it to describe the professional document of achievements, not starting something up again. The AI didn’t get that from the context, though I wish it did. If I could pronounce the word for AI, maybe it could then copy me. Also, I noticed contractions are tricky for the AI. The word here’s was pronounced as a 2 syllable word and the only way I could fix this was to go back and remove the contraction.