With speech-based addressing, instead of entering the mailbox number of a person or spelling the name using the telephone keypad, you can simply say the name. Outside callers, too, can say a name instead of entering an extension. The Basic Speech Auto Attendant recognizes first name-last name combinations (for example, William DuBois). To improve recognition accuracy in cases where the correct pronunciation of a name does not follow the pronunciation rules of the primary language, administrators can enter the name phonetically (for example, doobwah for Dubois).
If the confidence level of the speech recognition result is below a certain threshold, the Basic Speech Auto Attendant prompts the caller to confirm the name. The confirmation dialog then uses text-to-speech (TTS) to say the name and prompts the caller again to confirm the name. The caller is given several opportunities to confirm the name, enter an extension number, or the spelling of the name.