How to Use the Virtual Assistant Speech Commands?

Wait for Silence Before Saying your Command

Your assistant supports barge-in so you can interrupt a system voice prompt or text-to-speech, but for better speech recognition results, it is recommended to wait for a pause before saying your command. When reading e-mails and appointments, the assistant pauses shortly after reading the message header to allow you to say your command.

Use Speech Commands in their Context

Speech commands are spoken words or sentences that are part of a dialog expected by your virtual assistant. Even if the assistant can recognize a large vocabulary, specific commands are expected at specific dialog levels. The speech recognition technology parses the sentences you say and tries to recognize the commands expected at that stage of the dialog. For example, at the main menu level, you can say CALL CONTACT but you cannot say NEXT which applies only while listening to items in a list such as e-mails, appointments or voice mails.

Simple Speech Commands (Directed Dialog)

You can use speech for single action commands such as CALL CONTACT, DIAL NUMBER or SEND NEW MESSAGE. With single action commands, your assistant may initiate a dialog to ask for additional information to complete the command. As an example, if you say CALL CONTACT, it will ask: What is the name of your contact? , before suggesting the telephone number where it will call the selected contact. These Directed Dialog sessions enable the assistant to collect all the information it needs to execute your command.

Complex Speech Commands (Mixed Initiative)

Your assistant supports multi-part speech commands that include more than one information. For example, if you say CALL John Smith on his MOBILE PHONE, your assistant will understand the following set of speech commands and information: CALL+ John Smith+ MOBILE PHONE. This Mixed Initiative technology enables users to issue complete multi-part commands that sound like natural speech.

Select Contacts using Speech

Selecting contacts is much easier using speech than DTMF keys. You simply say the name of the contact. Your assistant expects you to say the phonetic name or the full name starting with the contact's first name than his/her last name. For example, to select John Smith, say JOHN SMITH . If there are more than one contact that sound like John Smith, you will be asked to confirm/select your contact from the list of possible matches.

Speak at Anytime (BARGE-IN)

You can speak your commands at anytime while messages are being played. The message being played is interrupted to enable the assistant to hear and understand your command. This technology is called Barge-In since it allows you to interrupt the assistant simply by speaking.

Speech Navigation Tips

Here are some tips to help you achieve the best speech recognition results over the phone:

Speak during silence/pauses between voice prompts for a better barge-in. Your assistant may not be able to understand your commands if it is already reading text or playing audio prompts.
Speak at a natural pace (not too fast and not too slow).
You don't have to speak loudly. You may achieve better results if you speak normally (not too loud and not too low).
Repeat your command if the assistant does not understand you the first time.
Switch to DTMF mode if you need privacy or if you are in a noisy environment.

Articles in this section