1-8
Cisco Personal Assistant 1.4 Installation and Administration Guide
OL-4590-03
Chapter 1 Planning for Personal Assistant
Understanding the Personal Assistant Server and Speech-Recognition Server

Access to Personal Assistant

The following sequence illustrates the processes involved when users access Personal Assistant:
1. Users dial the Personal Assistant number (the extension of the CTI route-point configured for
Personal Assistant in Cisco CallManager).
2. The call is routed to the first available media port. If no port is available, the call is rejected and the
caller hears a busy signal.
3. If a media port is available, the speech channel opens, and Personal Assistant plays a welcome
prompt. Users can then dial other users by name or access voice mail.
If speech resources are not available, Personal Assistant opens a non-speech session with the user,
enabling the user to use the phone keypad for touch-tone dial-by-name.
4. If users use dial-by-name to call another user:
a. User says: “Call ‘John Smith’.”
b. The command is received and processed by the speech-recognition server.
c. The user name is passed to Personal Assistant, and Personal Assistant retrieves user
information from the LDAP directory.
d. Based on the confidence level, Personal Assistant either prompts the user, or transfers the call.
Understanding the Personal Assistant Server and Speech-Recognition Server
Personal Assistant has a modular structure, which allows you to install various pieces of the product on
different server platforms. This provides a fault-tolerant redundant structure that you need for ensuring
that the Personal Assistant system is always available to your users.
The following sections provide information about the Personal Assistant servers and their interactions
with each other:
• Speech-Recognition Server, page 1-8
• Personal Assistant Server, page 1-9
• License and Resource Managers, page 1-9
• Interactions of the Personal Assistant Server and Speech-Recognition Server, page 1-10

Speech-Recognition Server

If you plan to implement any of the speech-enabled features of Personal Assistant, you must install at
least one speech-recognition server. During installation, you are given the option to install a
speech-recognition server, and you can either install it on the same system as the Personal Assistant
server or on a separate one.
The speech-recognition server prompts users and recognizes user responses based on a pre-defined
grammar, which includes a list of words and possible sequences in which a user can state them. During
a call to Personal Assistant, you can ask for help and Personal Assistant will tell you the available
commands. Also, the Help included with the Personal Assistant end-user interface defines this grammar
by using examples of valid prompts and responses.