Diagram of a VoiceXML Call

The following diagram shows a simplified representation of a VoiceXML call in the Unified Communications Managed API (UCMA) 3.0.

Call to a VoiceXML-Based IVR Application

  1. A person begins the call by entering the number for a service that they access by phone, for example a bank.

  2. The call arrives at a SIP server, such as the Microsoft Communications Server.

  3. The number is associated with a UCMA 3.0 Endpoint, and the call is directed to this endpoint.

  4. The hosting .NET application listens for calls at the UCMA 3.0 endpoint, and takes calls as they arrive.

  5. When the .NET application takes a call, it launches an instance of the Browser. The application launches a separate Browser instance for each call that it answers.

  6. The Browser connects to audio devices in the UCMA 3.0 endpoint that allow the Browser to present audio to the caller or collect audio from the caller as instructed by the VoiceXML application.

  7. The VoiceXML Interpreter in the Browser processes the markup language in the VoiceXML application. The VoiceXML application may be on a web server, or may be hosted within UCMA 3.0.

  8. The Browser communicates with the caller as instructed by the VoiceXML application. To present information to the caller, the Browser plays recorded audio files or verbalizes synthesized speech. To collect information from the caller, the Browser recognizes or records the caller's speech or keypad touch tones. The VoiceXML application cannot communicate directly with the caller.