Add CSI support for %speaker's names
