Platform SDK: Agent

Use the Active Voice

When using speech output to provide directive information or to elicit a user response, use the active voice and clearly specify the user's expected action. The following example illustrates the differences:

Let me repeat your number. No user action
The number will be repeated. Passive voice, no user action
Listen while the number is repeated. Passive voice
Listen to the repetition. Best choice

In addition, construct your output to unfold the key information at the end of the phrase as shown in the following examples:

Instead of: "Is three the next digit?" Use: "Is the next digit three?"
Instead of: "Click OK to begin." Use: "To begin, click OK."
Instead of: "Say 'Done' to complete your order." Use: "To complete your order, say 'Done.'"