It would be incredibly helpful if the system could better recognize and learn individual voices to reduce frequent mislabeling of speakers. Additionally, if a speaker’s name is corrected or updated in the transcript, that change should automatically apply across the entire transcript to maintain consistency.