Descript has the power to read and understand a transcript, yet filler word removal is totally one-size-fits-all. It would be great if it could scan the script on a per-speaker basis and suggest the removal of prevalent filler words that it's actually seeing.