Table Of Contents
Application Administration
Creating a New Application
Setting Up DNIS to the Application
Recognition Settings
Disambiguation Settings
Main Prompts Settings
Day and Night Settings
Weekend and Holiday Settings
Voicemail Setting
Name Collect Settings
Test Settings
Grammar Setup
Holiday Setup
Application Administration
This chapter describes the tasks required to set up Speech Connect applications, and to configure application parameters.
See the following sections:
•
Creating a New Application
•Setting Up DNIS to the Application
•Recognition Settings
•Disambiguation Settings
•Main Prompts Settings
•Day and Night Settings
•Weekend and Holiday Settings
•Voicemail Setting
•Name Collect Settings
•Test Settings
•Grammar Setup
•Holiday Setup
Creating a New Application
Revised May 2009
The system is shipped with a default application that contains preset values that you can use as your primary application, or you can use the default application as a template to create new applications. If multiple phone numbers are routed to the system, having multiple applications allows you to create a customized auto-attendant for each number. For example, you can have one application that is used for customers and other external callers, and another application for employees.
Note The Application Setup parameters are preconfigured with default settings that provide for optimal performance under normal usage. This chapter describes how you can modify the configuration settings if necessary to improve recognition or to change the way that the application interacts with callers.
To Create a New Application
Step 1 Click Config and then click Application. The Application Setup page appears with the default application settings.
Step 2 At the bottom of the page, click Create New From This Application.
Step 3 Enter the name of the new application, and click OK.
Note Application names must be unique and are not case sensitive. For example, if you have an application named "internal," you cannot create another application named "Internal."
Application names must not contain spaces. At this time, the administration interface erroneously allows spaces in application names. DNIS numbers routed to an application with spaces in the name will not answer.
Setting Up DNIS to the Application
Revised May 2009
By default, all phone numbers routed to the Speech Connect system will call the default application. If you have created a new application, the DNIS Setup page allows you to associate a phone number with the new application.
To Set Up DNIS to the Application
Step 1 In the Speech Connect interface, click Config and then click DNIS. The DNIS Setup page appears.
Step 2 Click Add. The DNIS Setup fields appear.
Step 3 Enter values into the fields, as applicable.
The DNIS Setup fields are described in Table 6-1.
Step 4 To update all values, click Update. The numbers will route to the specified applications.
Table 6-1 DNIS Setup Fields
Field
|
Description
|
Values
|
Number
|
Specifies the phone number or extension that is called to reach the application. You can set up multiple numbers to call one application.
|
Up to 20 characters
|
Application Name
|
Selects the application name to associate with the number.
|
Any valid application name
Note At this time, application names cannot contain space characters.
|
Delete
|
Deletes the selected number from DNIS setup. If the number is still routed to the system, it will call the default application.
|
•Unchecked—Do not delete the number.
•Checked—Delete the number.
|
Recognition Settings
Revised February 2008
If a caller speaks a name and Speech Connect finds multiple entries for that name, the system presents choices to the user and allows the user to confirm or deny each choice or to cancel the session. The Recognition settings allow you to tune the behavior of the system and its interaction with users to present and confirm names for recognition.
Recognition settings include:
•Name Directory—Controls the name response process in which the system either auto-accepts the name and transfers the call or interacts with the user by presenting a list of names to the caller.
•Confirm—Controls the confirmation process and how the system will recognize a yes or no response from the caller.
•Cancel—Controls how the system will recognize a cancel command from the caller.
The Recognition settings are preconfigured with defaults that provide for optimal recognition performance under normal usage. The settings are provided only if needed to improve recognition or change the way that the recognition engine interacts with callers. For example, you may want to configure different recognition settings for an external calling application than for an internal calling application.
To Configure Recognition Settings
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. The Recognition Settings appear on the Application Setup page.
Note If you have not created other applications, the default application settings page appears. If you have created other applications, a list of applications appears. Double-click the application you want to modify.
Step 2 Enter values into the fields, as applicable.
The Recognition Settings fields are described in Table 6-2.
Step 3 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-2 Recognition Settings Fields
Field
|
Description
|
Values
|
Name Directory: Min Confidence
|
Determines the confirmation range for name responses: the application confirms any result that falls between this score and Max Confidence. Everything below this score will be automatically rejected by the application. If the minimum confidence equals the Accept confidence, then all utterances are automatically accepted by the application.
|
Range: 0.0 to 0.99
Default: 0.05
|
Name Directory: Max (Accept) Confidence
|
Used to determine the auto-accept range for name responses. The application automatically accepts any result above the Accept confidence score.
|
Range: 0.0 to 0.99
Default: 0.6
Note Must be greater than or equal to Min Confidence
|
Name Directory: End Pointer
|
Specifies the length of required silence (in seconds) following user speech before the speech recognizer finalizes a result at the main prompt.
|
Range: 0 to 9.99
Default: 1.2
|
Name Directory: Number of Retries
|
The number of times the caller is allowed to retry a name response at the main prompt before the application transfers the caller to the failed operator, or dismisses the call if the failed operator is not configured.
|
Range: 1 to 10
Default: 5
|
Name Directory: Barge In
|
Specifies whether it is possible to barge in on prompts during any confirmations (disambiguation).
|
On = Enabled
Off = Disabled
|
Confirm: Min Confidence
|
Specifies the tuning cutoff point for recognizing a yes or no response during disambiguation. Any utterance recognized as yes or no with a confidence higher than this value will be acted on.
|
Range: 0 to 0.99
Default: 0.45
|
Confirm: End Pointer
|
Specifies the length of required silence (in seconds) following user speech before the speech recognizer finalizes a result while recognizing Confirm commands.
|
Range: 0 to 9.99
Default: 0.5
|
Confirm: Number of Retries
|
Specifies the number of times the caller is allowed to retry a yes or no response at the confirmation prompt before the application transfers the caller to the failed operator, or dismisses the call if the failed operator is not configured.
|
Range: 1 to 10
Default: 3
|
Confirm: Barge In
|
Specifies whether the caller is allowed to barge in when being asked for a yes or no response.
|
On = Enabled
Off = Disabled
|
Cancel: Min Confidence
|
Determines the cutoff point for recognizing the Cancel command. Any utterance recognized as Cancel with a confidence higher than this value will result in the cancel being recognized and acted on.
|
Range: 0 to 0.99
Default: 0.55
|
Cancel: End Pointer
|
Specifies the length of required silence (in seconds) following user speech before the speech recognizer finalizes a result while recognizing Cancel commands.
|
Range: 0 to 9.99
Default: 0.5
|
Grammar Package
|
Selects the name of the grammar package that is used by the application.
|
Any valid grammar package name up to 40 characters
|
ASR Time Out
|
Specifies the timeout value in seconds used during voice recognition.
|
Range: 0 to 20
Default: 5
|
Cancel Call Time Out
|
Specifies the timeout value in seconds used during cancel detection.
|
Range: 0.0 to 10
Default: 2
|
# of Visits to Main Menu
|
Specifies the total number of times a caller is allowed to visit the main prompt level during a single call. Callers who exceeds this limit are transferred to the Fail Operator.
|
Range: 1 to 100
Default: 25
|
Speed vs. Accuracy
|
Controls the desired balance between speed and accuracy. Higher values improve recognition but may affect recognition speed. Values lower then the default setting can result in low recognition and speed.
|
Range: 0.0 to 1.0
Default: 0.5
|
Disambiguation Settings
Revised February 2008
If a caller speaks a name for which multiple entries are found in the database, the system provides choices to the caller to isolate the intended party. The Disambiguation settings allow you to select the preferred tiebreaker method and control how the recognition engine interacts with the caller during the disambiguation process.
To Configure Disambiguation Settings
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. Navigate to the Disambiguation section.
Step 2 Enter values into the fields, as applicable.
The Disambiguation fields are described in Table 6-3.
Step 3 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-3 Disambiguation Fields
Field
|
Description
|
Values
|
Preferred Tie Breaker 1
|
Specifies the value of the Tiebreaker1 field that is used to sort the hypotheses list. All entries having this value are presented to the caller before any others.
|
Any valid employee Tiebreaker 1
|
Disambiguation Range
|
Specifies the range of hypotheses confidence scores that are considered ambiguous. When the disambiguation range differs between a hypothesis and subsequent hypotheses, the application will consider these hypotheses ambiguous.
|
Range: 0 to 0.99
Default: 0
|
# of Choices
|
Sets the maximum number of disambiguation choices to present to the caller. The subsequent choices will be presented only if the confidence score is above the minimum.
|
Range: 1 to 20
Default: 8
|
# of ASR Results
|
Controls the maximum number of disambiguation hypotheses that the Recognition engine is asked to return.
|
Range: 1 to 20
Default: 10
Note This value must be greater than # of Choices.
|
Main Prompts Settings
The Main Prompts settings allow you to change the system main prompts. You can view the transcription text for the main, retry, Help, and transfer prompts.
To Configure Main Prompts
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. Navigate to the Main Prompts section of the page.
Step 2 Enter values into the fields, as applicable.
The Main Prompts fields are described in Table 6-4.
Step 3 You can change any of the main prompts. To view the list of available prompts with their prompt verbiage, click Modify.
Step 4 In the selection box, click each prompt name to view the prompt verbiage. Optionally, you can select the "CUST" prompt to upload your own custom prompt.
Step 5 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-4 Main Prompts Fields
Field
|
Description
|
Values
|
Main
|
Displays the currently selected name of the main system prompt. This is the prompt played by Speech Connect when it requests that you speak an employee name.
|
Valid prompt name
|
Retry
|
Displays the currently selected name of the main retry prompt. This is the prompt played by Speech Connect when it does not understand the name spoken by a caller.
|
Valid prompt name
|
Help
|
Displays the currently selected name of the main Help prompt. This is the prompt played by Speech Connect when a caller says "help" when prompted to speak a name.
|
Valid prompt name
|
Transfer
|
Displays the currently selected name of the main transfer prompt. This is the prompt played by Speech Connect prior to a caller being transferred to their requested destination.
|
Valid prompt name
|
Day and Night Settings
The Day settings allow you to control the following during day operation:
•The greeting prompt played before the Main prompt (the default is blank)
•Operator phone number
•Failover phone number
Optionally, you can configure the welcome prompt, operator phone number, and fail phone number for evenings when the office is closed.
To Configure Day and Night Settings
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. Navigate to the Day and Night sections of the page.
Step 2 Enter values into the fields, as applicable.
The Day and Night fields are described in Table 6-5.
Step 3 You can change the Welcome prompt. To view the list of available prompts with their prompt verbiage, click Modify.
Step 4 In the selection box, click each prompt name to view the prompt verbiage. Optionally, you can select the "CUST" prompt to upload your own custom prompt.
Note If no verbiage is associated with a prompt, it is a blank prompt. A blank Welcome prompt allows the Main prompt to play without a specific greeting.
Step 5 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-5 Day and Night Fields
Field
|
Description
|
Values
|
Day
|
Welcome Prompt
|
Shows the greeting prompt to play before the main prompt during hours covered by the Day Operator.
|
Valid prompt name
|
Play Phone Number
|
Specifies whether the called party phone number is played before transferring.
|
Yes—Play phone number
No—Do not play phone number
|
Operator Number
|
Sets the number to transfer a call to when the caller says "Operator."
|
Any valid phone number, or NULL for not configured
|
Fail Phone Number
|
Operator Failed Call Phone number is the phone number to transfer to when too many errors occur.
|
Any valid phone number, or NULL for not configured
|
Night
|
Start Time
|
Specifies the time in a 24-hour clock format at which to start night time.
|
HH:MM or NULL to disable
Default: NULL
|
End Time
|
Specifies the time in a 24-hour clock format at which to end night time.
|
HH:MM or NULL to disable
Default: NULL
|
Welcome Prompt
|
Sets the greeting prompt to play during the night hours.
|
Valid prompt name
|
Play Phone Number
|
Specifies whether the called party phone number is played before transferring.
|
Yes—Play phone number
No—Do not play phone number
|
Operator Number
|
Sets the number to transfer a call to when the caller says "Operator."
|
Any valid phone number, or NULL for not configured
|
Fail Phone Number
|
Sets the number to transfer to when too many errors occur.
|
Any valid phone number, or NULL for not configured
|
Route Failed Calls to Operator
|
Sets whether to route calls to the operator number.
|
Yes—Route failed calls to the operator
No—Do not route failed calls to the operator
|
Weekend and Holiday Settings
Optionally, you can configure the welcome prompt, operator phone number, and fail phone number for weekends and holidays when the office is closed.
Note Holidays are defined in the Holiday Setup page. You have to define the holidays to enable the holiday settings in the application. For more information about the Holiday Setup page, see the "Holiday Setup" section.
To Configure Weekend and Holiday Settings
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. Navigate to the Weekend and Holiday sections.
Step 2 Enter values into the fields, as applicable.
The Weekend and Holiday fields are described in Table 6-6.
Step 3 You can change the Welcome prompt. To view the list of available prompts with their prompt verbiage, click Modify.
Step 4 In the selection box, click each prompt name to view the prompt verbiage. Optionally, you can select the "CUST" prompt to upload your own custom prompt.
Step 5 Configure the remaining parameters in the Application Setup screen, and click Update.
Note The Weekend and Holiday settings apply only during the actual weekend or holiday.
Table 6-6 Weekend and Holiday Fields
Field
|
Description
|
Values
|
Weekend
|
Use Weekend Settings
|
Specifies whether weekend settings are used.
|
Yes—Enable
No—Disable
|
Welcome Prompt
|
Shows the greeting prompt to play before the main prompt during weekend hours.
|
Valid prompt name
|
Play Phone Number
|
Specifies whether the called party phone number is played before transferring.
|
Yes—Play phone number
No—Do not play phone number
|
Operator Number
|
Sets the number to transfer a call to when the caller says "Operator."
|
Any valid phone number, or NULL for not configured
|
Fail Phone Number
|
Specifies the phone number to transfer to when too many errors occur.
|
Any valid phone number, or NULL for not configured
|
Route Failed Calls to Operator
|
Specifies whether to route calls to the operator number.
|
Yes—Route failed calls to the operator
No—Do not route failed calls to the operator
|
Holiday
|
Use Holiday Settings
|
Specifies whether holiday settings are used.
|
Yes—Enable
No—Disable
|
Welcome Prompt
|
Sets the greeting prompt to play during the holiday.
|
Valid prompt name
|
Play Phone Number
|
Specifies whether the called party phone number is played before transferring.
|
Yes—Play phone number
No—Do not play phone number
|
Operator Number
|
Specifies the number to transfer a call to when the caller says "Operator."
|
Any valid phone number, or NULL for not configured
|
Fail Phone Number
|
Specifies the number to transfer to when too many errors occur.
|
Any valid phone number, or NULL for not configured
|
Route Failed Calls to Operator
|
Specifies whether to route calls to the operator number.
|
Yes—Route failed calls to the operator
No—Do not route failed calls to the operator
|
Voicemail Setting
You can enter the phone number of your voice mail system. This setting will transfer callers to the voice mail system when they say, "Voice Mail."
To Set the Voicemail Phone Number
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. Navigate to the Voicemail section.
Step 2 Enter values into the field, as applicable.
The Voicemail field is described in Table 6-7.
Step 3 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-7 Voicemail Field
Field
|
Description
|
Values
|
Voicemail Number
|
Specifies the phone number to transfer to when a caller says "Voice Mail."
|
Any valid phone number, or NULL for not configured
20 characters (max)
|
Name Collect Settings
If you enable Name Collect, with an optional password, the system will prompt users to record their name if no recorded name file exists for that user.
To Configure Name Collect
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears. Navigate to the Voicemail section.
Step 2 Enter values into the fields, as applicable.
The Name Collect fields are described in Table 6-8.
Step 3 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-8 Name Collect Fields
Field
|
Description
|
Values
|
Enable
|
Specifies whether you want Name Collect processing for this application.
|
Yes—Enabled
No—Disabled
|
Password
|
Specifies the password a caller must enter during Name Collect before being allowed to record his or her name.
|
Up to 20 numeric digits
Blank = Disables password protection
|
Test Settings
Revised February 2008
You can use the Test Settings feature to easily determine whether Speech Connect recognition is working. If an automated test calls into Speech Connect and keys in the test ID, the system will respond by playing back the digits 987654321.
Note DTMF Tones may not be heard on an IP phone registered to Cisco Unified Communications Manager Version 5.x unless the "Disable RFC2833" option in the IP phone settings is set. Setting this option will force an MTP to be used, which allows translation of the RFC2833 events from Speech Connect into OOB DTMF for playback on the IP phone. This should be set for testing purposes only. Forcing the use of an MTP can lead to voice quality issues.
To Configure Test Settings
Step 1 In the Speech Connect interface, click Config and then click Application. The Application Setup page appears.
Step 2 Navigate to the Test section.
Step 3 In the Test ID field, enter digits that trigger the automated test.
The Test ID field is described in Table 6-9.
Step 4 Configure the remaining parameters in the Application Setup screen, and click Update.
Table 6-9 Test Field
Field
|
Description
|
Values
|
Test ID
|
Used to trigger an automated test. The system plays back a series of DTMF tones (987654321) for an auto-dialer test system to determine whether the application is running.
|
Up to 20 numeric digits
|
Grammar Setup
Revised July 2008
The Grammar Setup page allows you:
•Create a new grammar
•Set update parameters for each grammar
•Rebuild each grammar to include any updates to the employee and department records
•Create TTS recordings for employees who do not have a recorded name file as well as any new departments and tiebreakers. TTS processing runs in the background, and administrator control is returned.
After you create a new grammar, you can associate it with one of your calling applications. A default grammar called "Company" is set up to include all employees and departments in your database. Typically you would associate grammar containing all employees and departments with an internal calling application that allows your employees to reach every company extension. You can also create a grammar (called, for example, "External" or "Customers") that does not include all employees or departments, because you may not want external callers to reach certain employees or departments. If you set the List in External Directory field in an employee record to "no," and set the grammar to not include all employees, that employee will not be included in the grammar.
When the grammar is rebuilt through Grammar Setup, the TTS process is launched in the background. The TTS process generates a file containing a list of the employees who do not have a spoken name or TTS recording. Then the TTS tool takes that list and generates TTS recordings for those employees as well as any new departments and tiebreakers. You can also schedule the grammar rebuild and TTS processes to run periodically (by clicking Config > Scheduled Tasks). For more information about Scheduled Tasks, see the "Scheduling Names Tuning Tasks" section on page 8-2.
To Set Up Grammars
Step 1 In the Speech Connect interface, click Config and then click Grammar. The Grammar Setup page appears.
Step 2 Enter values into the fields, as applicable.
The Grammar Setup fields are described in Table 6-10.
Step 3 Click Update to update all fields.
Step 4 To rebuild the grammar, click Rebuild Now. The Grammar is rebuilt, and TTS processing is launched in the background.
Caution Rebuilding the grammar will take the application out of service. Be sure to schedule a grammar rebuild during off peak hours.
Table 6-10 Grammar Setup Fields
Field
|
Description
|
Values
|
Name
|
Name of this grammar.
Note Grammar names must be unique and are not case sensitive. For example, if you have a grammar named "company," you cannot create another grammar named "Company."
|
Up to 40 alpha characters
|
Language
|
Select the language from the list.
|
US English
|
Include all Employees
|
Specifies whether all employee records are included in the grammar. If set to "Yes," all employee records will be included in grammar. If set to "No," employees who are configured with List in External Directory set to "No" will not be included in the grammar.
|
Yes—Include all employee records
No—Include only the employees who are configured to be listed in the external directory
|
Include All Departments
|
Indicates that a department grammar should be created.
|
Yes—Create a department grammar
No—Do not create a department grammar
|
Delete
|
Deletes the selected grammar.
|
Unchecked: Do not delete
Checked: Delete
|
Holiday Setup
You can specify dates for during which your application will run in holiday mode. After you enable holiday mode in your application, the prompts and settings for holiday mode will be used during those dates.
To Set Up Holidays
Step 1 In the Speech Connect interface, click Config and then click Holiday. The Holiday Setup page appears.
Step 2 Enter values into the fields, as applicable.
The Holiday Setup fields are described in Table 6-11.
Step 3 To select a holiday from the calendar, double-click the Date field, navigate to the month and day, and click the date. The date appears in the field in MM/DD/YYYY format.
Step 4 Click Update to update all fields.
Note Holiday settings are by default disabled in your application. To enable holiday settings, click Config, and then click Application. In your application, navigate to the Holiday section and enable the Use Holiday Settings option.
Table 6-11 Holiday Setup Fields
Field
|
Description
|
Values
|
Date
|
Allows you to enter a date or select a date from the calendar.
|
Select a date from the calendar or enter date in MM/DD/YYYY format
|
Description
|
Enter a description of the holiday.
|
Up to 40 characters
|
Delete
|
Allows you to delete the selected holiday.
|
Unchecked—Do not delete holiday
Checked—Delete holiday
|
