Components of a voice application
In a typical dialog, a caller may provide several different pieces of information. For example, consider a simple dialog from the PizzaTalk sample application:
|
Application |
[Nuance earcon] Hi, and thanks for calling PizzaTalk with speech recognition by Nuance. So. . . what size pizza would you like? |
|
Caller |
Medium. |
|
Application |
And what do you want on it? |
|
Caller |
Pepperoni, mushrooms, and green peppers. |
|
Application |
You want a medium pizza with pepperoni, mushrooms, and green peppers. Is that right? |
|
Caller |
Yes. |
|
Application |
Great! Now what’s your phone number? |
|
Caller |
605-555-1234. |
|
Application |
I heard 605-555-1234. Did I get that right? |
|
Caller |
Yes. |
|
Application |
Thanks. By the way, we’ve already got your address on file. Now do you want your order to be delivered within the next hour? |
|
Caller |
Yes. |
|
Application |
OK. I got a medium pizza with pepperoni, mushrooms, and green peppers to be delivered to 1380 Willow Road. Thanks for calling PizzaTalk. Ciao! |
Here, the caller is prompted for the pizza size, up to three different toppings, and a phone number—as well as confirmations at different points in the dialog.
From the caller’s point of view, all of this happens in a smooth, natural dialog. However, behind the scenes, Voice Platform is using several different VoiceXML files, grammars, Java Server Pages, and prerecorded prompts.
