It would be awesome if there was a backward arrow/option when identifying speakers using AI for transcription. I find that if I accidentally label a speaker, I have to close out of the box and start all over again. Would love to just toggle back a step and reset that speaker.