Avaya Speech Applications Builder User Manual

Download

Page 1

Avaya Speech Applications Builder Configurator User Guide

May 15, 2004

Page 2

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 2 of 116

Notice

While reasonable efforts were made to ensure that the information in this document was complete and accurate at the time of printing, Avaya Inc. can assume no liability for any errors. Changes and corrections to the information in this document may be incorporated in future releases.

Preventing toll fraud

"Toll fraud" is the unauthorized use of your telecommunications system by an unauthorized party (for example, anyone who is not a corporate employee, agent, subcontractor, or person working on your company's behalf). Be aware that there may be a risk of toll fraud associated with your system and that, if toll fraud occurs, it can result in substantial additional charges for your telecommunications services.

Avaya fraud intervention

If you suspect that you are being victimized by toll fraud and you need technical assistance or support, call Technical Service Center Toll Fraud Intervention Hotline at +1-800-643-2353 for the United States and Canada. For additional support telephone numbers, see the Avaya Web site:

http://www.avaya.com

Select Support, then select Escalation Lists. This Web site includes telephone numbers for escalation within the United States. For escalation telephone numbers outside the United States, select Global Escalation List.

Providing telecommunications security

Telecommunications security (of voice, data, and video communications) is the prevention of any type of intrusion to (that is, either unauthorized or malicious access to or use of) your company's telecommunications equipment by some party.

Your company's "telecommunications equipment" includes both this Avaya product and any other voice/data/video equipment that could be accessed via this Avaya product (that is, "networked equipment").

An "outside party" is anyone who is not a corporate employee, agent, subcontractor, or person working on your company's behalf. Whereas, a "malicious party" is anyone (including someone who may be otherwise authorized) who accesses your telecommunications equipment with either malicious or mischievous intent.

Such intrusions may be either to/through synchronous (timemultiplexed and/or circuit-based) or asynchronous (character-, message-, or packet-based) equipment or interfaces for reasons of:

• Use (of capabilities special to the accessed equipment)

• Theft (such as, of intellectual property, financial assets, or toll-facility access)

• Eavesdropping (privacy invasions to humans)

• Mischief (troubling, but apparently innocuous,

tampering)

• Harm (such as harmful tampering, data loss or alteration, regardless of motive or intent)

Be aware that there may be a risk of unauthorized intrusions associated with your system and/or its networked equipment. Also realize that, if such an intrusion should occur, it could result in a variety of losses to your company (including, but not limited to, human and data privacy, intellectual property, material assets, financial resources, labor costs, and legal costs).

Your responsibility for your company's telecommunications security

The final responsibility for securing both this system and its networked equipment rests with you, an Avaya customer's system administrator, your telecommunications peers, and your managers. Base the fulfillment of your responsibility on acquired knowledge and resources from a variety of sources, including, but not limited to:

• Installation documents

• System administration documents

• Security documents

• Hardware-/software-based security tools

• Shared information between you and your peers

• Telecommunications security experts

To prevent intrusions to your telecommunications equipment, you and your peers should carefully program and configure:

• Your Avaya-provided telecommunications systems and their interfaces

• Your Avaya-provided software applications, as well as their underlying hardware/software platforms and interfaces

• Any other equipment networked to your Avaya products.

Part 15: Class A Statement

For the MCC1, SCC1, G600, and CMC1 Media Gateways:

Note: This equipment has been tested and found to comply with the limits for a Class A digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide reasonable protection against harmful interference when the equipment is operated in a commercial environment. This equipment generates, uses, and can radiate radio frequency energy and, if not installed and used in accordance with the instruction manual, may cause harmful interference to radio communications. Operation of this equipment in a residential area is likely to cause harmful interference, in which case the user will be required to correct the interference at his own expense.

Part 15: Class B Statement

For the G700 Media Gateway:

Note: This equipment has been tested and found to comply with the limits for a Class B digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide reasonable protection against harmful interference in a residential installation. This equipment generates, uses, and can radiate radio-frequency energy and, if not installed and used in accordance with the instructions, may cause harmful interference to radio communications. However, there is no guarantee that interference will not occur in a particular installation. If this equipment does cause harmful interference to radio or television reception, which can be determined by turning the equipment off and on, the user is encouraged to try to correct the interference by one or more of the following measures:

• Reorient the receiving television or radio antenna where this may be done safely.

• To the extent possible, relocate the receiver with respect to the telephone equipment.

• Where the telephone equipment requires AC power, plug the telephone into a different AC outlet so that the telephone equipment and receiver are on different branch circuits.

• Consult the Dealer or an experienced radio/TV technician for help.

Page 3

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 3 of 116

Canadian Department of Communications (DOC) Interference Information

For MCC1, SCC1, G600, and CMC1 Media Gateways:

This Class A digital apparatus complies with Canadian ICES-

003.

Cet appareil numérique de la classe A est conforme à la norme NMB-003 du Canada.

For the G700 Media Gateway:

This Class B digital apparatus complies with Canadian ICES-

003.

Cet appareil numérique de la classe B est conforme à la norme NMB-003 du Canada.

This equipment meets the applicable Industry Canada Terminal

Equipment Technical Specifications. This is confirmed by the registration number. The abbreviation, IC, before the registration number signifies that registration was performed based on a Declaration of Conformity indicating that Industry Canada technical specifications were met. It does not imply that Industry Canada approved the equipment.

Japan

For the MCC1, SCC1, G600, and CMC1 Media Gateways:

This is a Class A product based on the standard of the Voluntary Control Council for Interference by Information Technology Equipment (VCCI). If this equipment is used in a domestic environment, radio disturbance may occur, in which case, the user may be required to take corrective actions.

For the G700 Media Gateway:

This is a Class B product based on the standard of the Voluntary Control Council for Interference by Information Technology Equipment (VCCI). If this equipment is used in a domestic environment, radio disturbance may occur, in which case, the user may be required to take corrective actions.

Part 15: Personal Computer Statement

This equipment has been certified to comply with the limits for a Class B computing device, pursuant to Subpart J of Part 15 of FCC Rules. Only peripherals (computing input/output devices, terminals, printers, etc.) certified to comply with the Class B limits may be attached to this computer. Operation with noncertified peripherals is likely to result in interference to radio and television reception.

Part 68: Answer-Supervision Signaling

Allowing this equipment to be operated in a manner that does not provide proper answer-supervision signaling is in violation of Part 68 rules. This equipment returns answer-supervision signals to the public switched network when:

• answered by the called station,

• answered by the attendant, or

• routed to a recorded announcement that can be

administered by the CPE user.

This equipment returns answer-supervision signals on all direct inward dialed (DID) calls forwarded back to the public switched telephone network. Permissible exceptions are:

• A call is unanswered.

• A busy tone is received.

• A reorder tone is received.

DECLARATIONS OF CONFORMITY

US FCC Part 68 Supplier’s Declaration of Conformity (SDoC)

Avaya Inc. in the United States of America hereby certifies that the Avaya switch equipment described in this document and bearing a TIA TSB-168 label identification number complies with the Federal Communications Commission’s (FCC) Rules and Regulations 47 CFR Part 68, and the Administrative Council on Terminal Attachments (ACTA) adopted technical criteria.

Avaya further asserts that Avaya handset equipped terminal equipment described in this document complies with Paragraph

68.316 of the FCC Rules and Regulations defining Hearing Aid Compatibility and is deemed compatible with hearing aids. Copies of SDoCs signed by the Responsible Party in the US can be obtained by contacting your local sales representative and are available on the following Web site:

http://www.avaya.com/support

All Avaya switch products are compliant with Part 68 of the FCC rules, but many have been registered with the FCC before the SDoC process was available. A list of all Avaya registered products may be found at:

http://www.part68.org/

by conducting a search using "Avaya" as manufacturer.

European Union Declarations of Conformity

Avaya Inc. declares that the equipment specified in this document bearing the "CE" (Conformité Europeénne) mark conforms to the European Union Radio and Telecommunications Terminal Equipment Directive (1999/5/EC), including the Electromagnetic Compatibility Directive (89/336/EEC) and Low Voltage Directive (73/23/EEC). This equipment has been certified to meet CTR3 Basic Rate Interface (BRI) and CTR4 Primary Rate Interface (PRI) and subsets thereof in CTR12 and CTR13, as applicable. Copies of these Declarations of Conformity (DoCs) signed by the Vice President of R&D, Avaya Inc., can be obtained by contacting your local sales representative and are available on the following Web site:

http://www.avaya.com/support

TCP/IP facilities

Customers may experience differences in product performance, reliability, and security, depending upon network configurations/design and topologies, even when the product performs as warranted.

Warranty

Avaya Inc. provides a limited warranty on this product. Refer to your sales agreement to establish the terms of the limited warranty. In addition, Avaya’s standard warranty language, as well as information regarding support for this product, while under warranty, is available through the following Web site:

http://www.avaya.com/support

Link disclaimer

Avaya Inc. is not responsible for the contents or reliability of any linked Web sites and does not necessarily endorse the products, services, or information described or offered within them. We cannot guarantee that these links will work all of the time and we have no control over the availability of the linked pages.

Page 4

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 4 of 116

Trademarks

Avaya is a trademark of Avaya Inc.

Insert all other Avaya Trademarks here, then delete this paragraph. DO NOT include other company’s trademarks.

All trademarks identified by the ® or ™ are registered trademarks or trademarks, respectively, of Avaya Inc. All other trademarks are the property of their respective owners.

Ordering information: Avaya Publications Center

Voice: +1-207-866-6701

1-800-457-1764 (Toll-free, U.S. and Canada only)

Fax: +1-207-626-7269

1-800-457-1764 (Toll-free, U.S. and Canada only)

Write: Globalware Solutions

200 Ward Hill Avenue

Haverhill, MA 01835 USA

Attention: Avaya Account Manager

Web: http://www.avayadocs.com

E-mail: [email protected]

Order:

Avaya support

Avaya provides a telephone number for you to use to report problems or to ask questions about your contact center. The support telephone number is1-800-242-2121 in the United States. For additional support telephone numbers, see the Avaya Web site:

http://www.avaya.com

Comments

To comment on this document, send e-mail to

[email protected].

Acknowledgment

This document was written by the CRM Information Development group.

Page 5

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 5 of 116

Contents

Contents ............................................................................................................................................. 5

About this document ......................................................................................................................... 7

Audience 7

Related documents include:

• Speech Applications Builder Release Notes

• Speech Applications Builder Installation Guide

• Configuring Speech Applications Builder for Deployment

• JBOSS Configuration and Deployment Guide

• Speech Applications Builder Component List

• Speech Applications Builder Component Developer’s Guide

Page 8

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 8 of 116

System requirements

The minimum system requirements for running Speech Applications Builder are as follows:

Hardware requirements

Minimum speed required for platform 500MHz

Minimum memory required for platform 512MB RAM

Software requirements

Speech Applications Builder Voice Applications Configurator Tool

Operating System

Speech Applications Builder has been verified to run under Windows 2000.

However, it should also be capable of running under any operating system that supports Java™ 1.4.1.

Page 9

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 9 of 116

Introduction to Speech Applications Builder

Some terminology

The Speech Applications Builder (SAB) Configurator is a tool for building, testing, deploying and updating voice-driven automated telephone services, such as Telephone Banking or Catalogue Ordering.

An automated telephone service consists of one or more applications.

An application includes one or more interactions with the caller. These interactions are called dialogs. The software that performs a dialog is called a dialog flow (or a dialog model).

The main dialog flow may contain one or more sub-dialog flows (and a sub-dialog flow may have sub flows of its own).

SAB dialog flows are built using pre-defined SAB components, and are created as a type of flow diagram, drawn using the SAB GUI. For a description of the SAB GUI, see page 19.

There are several types of SAB component:

• dialogs - which handle interactions with the caller and may include speech recognition

• processes - which retrieve data and perform data manipulation

• rules - which allow branching or decision making

• interrupt - which is used to provide an alternative dialog flow in response to a particular

recognition result, such as the caller saying “Help” or “Main Menu”

For an overview of how the components are used, see page 10.

For a List of Components, see page 108.

The SAB GUI

SAB dialogs are developed using the SAB GUI. This provides:

• Development facilities - which allow you to create a dialog by selecting and configuring

components and connecting them together in a call flow. The GUI also provides debugging tools.

• Search facilities - which allow you to search for components and prompt files, by name or

partial name.

• A file directory - which provides easy access to dialog flows and prompt files.

• A version control mechanism - which allows you to label and store earlier versions of a

dialog flow. Speech Applications Builder also provides a formal mechanism for migrating a dialog from the development phase (the tool environment) to the deployment phase (the

production environment) and then to the decommissioned phase (the decommission environment).

• Administrative functions.

For a description of the SAB GUI, see page 19.

Page 10

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 10 of 116

The structure of a dialog flow

A dialog flow is built by selecting and configuring pre-defined components and connecting them together in a call flow. There are three main types of component: dialogs, processes and rules. An instance of a component is called a step.

Figure 1. The different step icons (left to right): dialog, process, rule and Begin Dialog step

When you create a new dialog flow (see page 34), SAB automatically creates a Begin Dialog step and an End Dialog step.

Figure 2. The Dialog Modeler window showing the Begin Dialog step (orange) and the three

types of component: dialogs (blue), processes (yellow), and rules (black diamonds)

Page 11

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 11 of 116

Dialog steps

Generally speaking, dialog steps handle interactions with the caller. Figure 2 shows a series of dialog steps that

• receive the call, setting up variables to store the CLI, DDI, Port, call time and other information

• play a welcome message (such as Welcome to Acme loan center customer service)

• offer the caller a menu (To inquire about the status of your loan, please say “Status”. For all

other enquiries please say “Operator”’) and recognize his or her response

• ask the caller to say his or her loan number (What is your loan number?) and recognize the response

• ask for a “password” (For verification purposes, what are the last four digits of your Social Security Number?) and recognize the response

• deal with an invalid loan number (I’m sorry the loan number you provided is not a valid number. Please say your loan number now.)

... and so on.

All dialog components that perform speech recognition also handle

• confirmation (Was that 1234?), if required

• no input or silence (I’m sorry I did not hear anything. Please say it again)

• no match (I’m sorry I did not understand that. Please say it again)

• misrecognition

Process steps

Process steps handle data. Figure 2 shows a process step that accesses a database to retrieve the customer record associated with a particular loan number. Other SAB processes include:

• addition, subtraction, multiplication, division, total, exponentiation

• check lookup table

• capitalize, change case

• set variable, copy variable, insert new variable

• report call end, report call information

• send event code

and so on. The data being manipulated may be static, or derived from a recognition result, or retrieved from a database.

Connections and Rules

The flow of the dialog is created by drawing connections between the steps.

By default, each connection includes a decision point (these are the grey diamonds on some of the connections in Figure 2). If the dialog needs to branch, a rule can be defined on the decision point (these are the black diamonds in the Figure) and a second connection added.

Page 12

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 12 of 116

SAB rules are boolean. They include

• less than, more than, equal to

• text handling rules

• list handling rules

• looping

More complex boolean rules can be built using multiple rule components.

For example

If id = 10 AND name = "Jones”

can be built as shown in the Figure below, where the first operator is defined as id = 10 and the second as name = “Jones”.

Figure 3. A complex rule

For more information about defining rules, see page 64.

Triggers

Certain types of event - including loggable events and Java exceptions - can be handled using triggers. These are (typically short) alternative dialog flows that are run when the triggering event occurs.

There are five types of trigger:

• Exception - evaluated only when an exception occurs in the underlying code

• Before step validated - legacy feature. DO NOT USE

• Before step runs - evaluated before the step

• After step runs - evaluated after the step

• Unhandled control - evaluated if the path required is not defined in the dialog flow

Page 13

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 13 of 116

Triggers may be configured

• as part of a flow, in the Trigger Modeler window (see page 62). These triggers are used by default when any triggering event occurs in that flow, if the step in which the event occurs does

not specifically handle triggers.

• as part of a step, in the Edit Step dialog box (see page 63). Step-based triggers are specific to the particular instance of the component and take priority over any global triggers.

Prompts

The recorded or synthesized words played out to the caller are called prompts. Prompts are the automated system’s contribution to the “conversation”.

Some prompts ask the caller for information and require a reply before the dialog can continue. Some prompts give the caller information and do not require a reply. A prompt that does not require a reply is often called a message.

Part of the process of configuring a dialog component is defining the prompts it will use. These are typed into the dialog boxes provided by the component, which also provides a simple method for including variable data and for turning on Barge-in. (Barge-in allows the caller to interrupt the prompt before it has finished playing. The speech recognizer listens for speech during the prompt, stops the prompt playout when it detects a response, and performs the recognition.)

For example,

Typed

Variable dragged and dropped from

a list of previously defined variables

Typed

Command dragged

and dropped

Typed

Did you

say

<caller_data>

<BARGE-IN_ON>

Please say

yes or no.

Figure 4. The construction of a typical confirmation prompt

When the dialog flow is run, SAB sends the text of the prompt to the Text-To-Speech (TTS) synthesizer, which synthesizes a speech equivalent of the text and returns it to SAB to be played out

to the caller.

Synthesized speech is particularly useful during the development of an application, but for a deployed system many service providers prefer to use a recorded voice.

Recorded prompts are managed in the Prompts resource (see below and page 82), which allows you to them either by recording and editing them yourself or by specifying .wav files (which may be professionally recorded).

SAB also provides support for handling

• multiple languages

The list of languages handled is specified by the user in the Administration resource (see below, and page 93). This, in effect, sets up a separate directory structure for each language. When you define a prompt, you specify its language by selecting from a drop-down list, and SAB labels the file accordingly. This mechanism allows you to define multiple sets of prompts (each set in a different language) that use the same prompt names but are automatically stored in different directories. This, in turn, allows you to create a single application that uses multiple languages.

Page 14

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 14 of 116

Note that some languages have multiple variants or 'locales' (for example, US English, UK English) which are pronounced differently. When you specify a language with multiple locales, ensure that you use the complete language ID, including the correct locale, for example, “enUS” (for US English), or “en-GB” (for GB English).

• multiple personae

The list of personae handled is specified by the user in the Administration resource (see below, and page 93). This, in effect, sets up a separate directory structure for each persona. When you define a prompt, you specify its persona by selecting from a drop-down list, and SAB labels the file accordingly. This mechanism allows you to define multiple sets of prompts (each set using a different persona) that use the same prompt names but are automatically stored in different directories. This, in turn, allows you to create a single application that uses multiple personae. Persona is discussed on page 90.

The SAB database

SAB stores all components, all dialog flows, and all prompts in a database called the SAB database. A default database is created when the SAB Configurator is installed. New or updated components can be imported into the SAB Configurator via the Component Workbench, at any time, and SAB provides mechanisms for updating existing dialog flows in the Tools environment.

For more information about the Component Workbench, see page 92.

When the SAB Configurator is updated, both the Configurator and the SAB database are uninstalled. It is therefore essential to export all dialog flows to file (see page 77) before uninstalling SAB. The dialog flows can be re-imported after the Configurator has been upgraded.

Debugging facilities

The Speech Applications Builder provides facilities for analyzing and debugging a dialog flow. Dialog Analysis lists and describes every path through the selected dialog flow. The data it provides can be

used to build test cases.

Run/Debug allows you to “run” a dialog, telephone it, and watch the call flow as it happens. You can open an individual dialog step and check its configuration as you hear it being executed.

Version control

The SAB Configurator provides the following version control mechanisms:

• SAB automatically adds a version number (1.0) to a new dialog flow and provides both a copy with same version number facility (see page 74) and a copy with an incremented version number facility (see page 74). All versions of a dialog flow are stored in the SAB database,

and are displayed in the Dialogs tree, in the Dialog version index and following a search.

• SAB is a multi-user development environment. To prevent more than one developer simultaneously editing a dialog flow, SAB automatically opens it in a read-only version. The developer cannot edit the dialog flow until he or she checks it out (see page 37). Whilst a dialog flow is checked out, no other developer can edit it (though he or she may open another read only version). The dialog flow should be checked in when the developer has finished editing.

Page 15

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 15 of 116

• SAB is both a development environment and a runtime environment, and provides a formal mechanism for migrating a dialog from the development phase (the tool environment) to the deployment phase (the production environment) and then to the decommissioned phase (the decommission environment).

o When a dialog flow is in the tool environment, it can be edited.

o When the dialog flow has been migrated to the production environment it can be

viewed and executed but cannot be edited.

o When a dialog flow has been migrated to the decommission environment it cannot

executed. However, the decommissioned version can be copied (with an incremented version number), back into the tool environment and edited, and then re-deployed.

Page 16

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 16 of 116

Starting SAB

The Installation procedure is described in the Speech Applications Builder Installation Guide.

To launch SAB:

1. Either double-click on the SAB icon, or select

Start > Programs > Avaya Speech Applications Builder

Logging in

If this is the first time you have opened the Configurator, SAB will ask you to enter a License Key (see page 17).

Otherwise, SAB displays the User Authentication dialog box, which asks you to enter a UserID and Password.

For details of how to change your Password, see page 17.

The SAB Configurator is installed with the following defaults:

UserID: administrator Password: password

To log in:

1. Enter your User ID and Password.

2. Click OK.

Figure 5. User Authentication dialog box

Page 17

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 17 of 116

Entering the License Key

If this is the first time you have opened the Configurator, SAB asks you to enter a License Key.

To enter the License Key:

1 Enter or copy it into the dialog box.

Figure 6. License Key dialog box

2 Click OK.

Changing your password

To change your password:

1 Open the SAB Configurator.

2 Either click Change Password on the User Authentication dialog box (see page 16) or, from the

Configurator Menu bar, select File > Edit System Preferences.

3 Select User Preferences.

4 Click on Change Password. SAB displays the following dialog box.

Figure 7. Change Password

5 Enter your old password.

6 Enter your new password twice.

7 Click OK.

Page 18

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 18 of 116

Exiting SAB

To exit SAB, select File > Exit.

If you have made unsaved changes to a dialog flow, SAB will ask you if you want to save the dialog flow.

Figure 8. Save warning message

Page 19

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 19 of 116

The SAB GUI

This chapter provides a brief description of the SAB Graphical User Interface (GUI). The features discussed here are described in more detail at the appropriate places in the remainder of the manual.

The SAB GUI consists of

• a Menu bar and Tool bar, running along the top, which display both generic and context- sensitive options

• a Browser, on the left, which is a tool used to display, search for and manage dialogs and their related resources, and to perform system administration

• an area, on the right, that is primarily used to display the Dialog Modeler, which is used to create and manage dialog flows. This area is also used to:

o create and edit recorded prompts

o import and manage components

o perform system administration.

Figure 9. The SAB GUI

Page 20

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 20 of 116

The Menu bar

The Menu bar provides the following generic functions:

Dialog modeler

Used to configure the Dialog Modeler display.

User preferences

Used to change password..

Edit System Preferences

See page 105.

General preferences

Used to configure general features of the Configurator.

File >

Exit

Closes SAB.

Window

Used to toggle between instances of the Dialog modeler.

The Menu bar also provides a Dialog Model menu and a Prompt menu, when appropriate:

When the Dialogs tree is selected in the Browser

Add Dialog Model

Opens a new dialog flow window (or dialog model) that can then be checked out and used to develop a new dialog flow. See page 34.

Imports a Dialog Model from a File

Used to import an SAB file stored outside the SAB database. See page 79.

Dialog Model >

Exports the Dialog Model to a File

Used to save an SAB dialog flow file to a location outside the SAB database. See page 77.

When a dialog flow is selected in the Browser

Save

Saves the currently selected dialog flow in the SAB database. See page 77.

Note that dialog flows can also be exported to file.

Copy Dialog Model and a selected version

Creates a copy of the selected dialog flow file, labeled “Copy of”. See page 74.

Deletes the Selected Dialog Model Version

Deletes the selected dialog flow. See page 72.

Rename

Renames the selected dialog flow. See page 75.

Dialog Model >

Increments the Version of the Selected Dialog Version

Creates a copy of the selected dialog flow file with a new version number. See page 74.

Page 21

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 21 of 116

Add Prompt

Used to create a new recorded prompt. See page 82.

Import Wave Zip

Used to import a zip archive containing recorded prompts. See page 89.

Prompt >

Create Wave Extract

Used to create a zip file that contains prompts. See page

89.

The Tool bar

The Toolbar, which runs along the top of the SAB Configurator, provides short cuts to some of the options available from the Menu bar.

Generic buttons

Add Organization

Add ...

Opens a drop down menu of “create new item” options:

Organization Node

Environment

For example, a QA environment. This would be similar the Production environment, in that dialog flows could not be edited.

Permission Type

Add Dialog

Opens an Add New Dialog dialog box. See page 34.

Add Prompt

Opens an Add New Prompt dialog box. See page 82.

Edit System Preferences

Used to configure certain features of the SAB GUI. See page 105.

Page 22

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 22 of 116

Dialogs buttons

The following buttons are provided when the Dialogs tree is in use. The selection of buttons displayed is relevant to the task being performed.

Add a Dialog Model

Opens a new dialog flow window complete with Begin Dialog and End Dialog steps.

Imports a Dialog Model from a File

Exports a Dialog Model to a File

Saves the current Dialog Model (all versions)

The Dialog flow is saved to the SAB database.

Check out a Dialog Model

Opens an editable version of the dialog flow and locks out other users. Since SAB is a multi-user environment, checking in and checking out ensures that only one user can work on a dialog at any time.

Check in a Dialog Model

Closes the editable version of the dialog flow and makes it accessible to other users

Copy Dialog Model and selected version

Copies the selected version of a dialog flow, creating an exact copy labeled “copy”.

Increments the Version of the Selected Dialog Version

Copies the dialog and places copy in the Tool environment, incrementing the version number (for example 1.0 becomes 2.0)

Generate War File for Deployment

Deletes the Selected Dialog Model

Migrate Package

Migrates a dialog from one environment to another.

Decommission

Takes a dialog out of Production

Some of these options are also available from the pop-up Context menu displayed by right clicking on the dialog flow name in the Dialogs tree.

Page 23

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 23 of 116

Prompt buttons

The following buttons are provided when the Prompts resource is in use. The selection of buttons displayed is relevant to the task being performed.

Add (create) a prompt

Import a zipped prompt directory. See page 89.

Export a zipped prompt directory. See page 89.

Delete

Context menus

Many of the options provided on the Menu bar and Tool bar are also presented on pop-up Context menus, which can be displayed by

• right-clicking on an icon in the Browser

• right-clicking in the Dialog Modeler window.

The options presented in a pop-up menu are relevant to the task being performed.

Refresh Tree From Here

Refresh Tree From Here is available from the Browser pop-up menus at strategic points in the Dialogs and Prompt trees. Selecting the option refreshes all the child nodes of the selected node, performing any updates that have not occurred automatically.

Page 24

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 24 of 116

The Browser

The Browser is a tool used to display, search for and manage dialog flows and prompts, import components, and perform system administration. The Browser provides access to four sets of resources, called Dialogs, Prompts, Component Workbench and Administration.

Figure 10. The Browser, showing the Dialogs, Prompts, Component Workbench and

Administration resources

• Dialogs is used to manage dialog flows. See page 33.

• Prompts is used to manage prompts. See page 82.

• The Components Workbench is used to import new components and manage existing

components. See page 92.

• Administration is used to manage SAB. See page 93.

In some respects, the tree displayed in the Browser acts like a typical directory structure. For example, all Dialog flow files are displayed beneath the Dialogs node.

Figure 11. The Dialogs directory structure, expanded to show the stored dialog flows

Page 25

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 25 of 116

By default, the dialog flows are displayed in alphabetical order, but you can group and nest them as required by specifying a Classification (see page 34) when you create the dialog flow file.

Figure 12. The Dialogs directory structure, showing a dialog flow “Main Dialog” classified under

“Direct Debit”, which is itself classified under “Banking”.

If you click on a dialog flow icon, (for example, Main Dialog in the Figure above, which is, in effect, a leaf node), SAB displays the Dialog Model Version Index, shown in the Figure below, which lists all the versions of the dialog flow available, and you can open the version you require by clicking on the entry.

Figure 13. Part of the Dialog Model Index

Alternatively, you can expand the dialog flow in the tree, and open a particular version of the dialog flow by clicking on the version number.

All the trees work in a similar way. The Prompts tree displays a list of prompt folders, which can be expanded to display individual prompt files. These can be opened for editing by clicking in the file icon to display the prompt version index and then clicking on the version number required.

Page 26

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 26 of 116

Figure 14. The Prompts directory structure, expanded

Expanding the Component Workbench opens the Repository Maintenance window, which can be used to import new components. Expanding Administration displays the Configuration, Organization and Permissions options.

Search engine

In addition to its file directory and feature directory function, the tree also provides a Search engine that can be used to search for

• a dialog flow (see page 73)

• a prompt (see page 87)

To search for any entity:

1 Click on the appropriate node; for example, to search for a dialog flow, click on one of the Dialogs

nodes. (Note, however, that the Search engine is not available from a leaf node). Voice Runner displays its Search engine in the Dialog Modeler window.

2 Type the name, or partial name, of the entity you want to find into the Name field. If you have used

a partial name, ensure that the Partial Match box is also checked.

3 In the case of a dialog flow or a prompt, you can select the environment you want to search:

• Any environment - searches all environments; this is the default

• Tool - searches the Tool environment. This contains dialog flows that are under construction

(and related prompts)

• Production - searches the Production environment. This contains dialog flows that are deployed (and related prompts)

• Decommission - searches the Decommission environment. This consists of dialog flows that have been taken out of deployment (and related prompts).

4 Click Search. SAB displays a list of all matching entities in the Dialog Modeler window.

5 Click Show Details to display further details.

6 Click on the entity name to open it in the Dialog Modeler window.

Page 27

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 27 of 116

Thumbnail and zoom facilities

When a dialog flow is open in the Dialog Modeler window, the Browser also provides

• a thumbnail view of the entire dialog flow, outlining the area that is also displayed in the Dialog Modeler window with a red border

Figure 15. The Browser thumbnail view

Page 28

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 28 of 116

• a “zoomed” view of the dialog flow, which can be scrolled by moving the cursor around the Dialog Modeler window.

Figure 16. The Browser “zoomed” view

Page 29

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 29 of 116

The Dialog Modeler window

The Dialog Modeler window (on the right of the Configurator) is a workspace where you can build new dialogs and view or amend existing dialogs. It allows you to build, test, analyze and deploy dialog flows.

Figure 17. The Dialog Modeler window (at right hand side)

The Dialog Modeler window contains the following tabbed worksheets:

Worksheet

Function

Dialog Modeler

The default worksheet, which is used to create and edit dialog flows by “drawing”.

Trigger Modeler

Used to create alternative dialog flows to handle specified events occurring in this dialog flow.

Note that triggers can also be set up within the individual component steps and, if these exist, take precedence over the dialog-wide triggers.

See page 62.

Required Data

Used to set up the variables and arrays that are required by this sub-dialog flow. See page 68.

Page 30

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 30 of 116

Worksheet

Function

Output Data

Used to set up the variables and arrays used to pass data out of this sub-dialog flow. See page 68.

Global Settings

Used to configure certain global recognition parameters, including

• barge-in (see Glossary, page 114)

• default minimum confidence (see Glossary, page 114)

• no input timeout - defines the length of silence that must elapse before SAB

classifies the response as no input

• maximum errors allowed - defines the maximum number of times SAB attempts to obtain a valid response.

Dialog Analysis

Used to analyze paths through the dialog and check for unreachable Steps (steps that cannot be executed). See page 100.

Run/Debug

Used to test the dialog in the debugger test environment. See page 102.

Event Log

Logs “development events”, such as Save and Export.

Drawing facilities

The Dialog Modeler window provides a “canvas”, on which the dialog flow is drawn, and a set of drawing tools, accessed in either of the following ways:

• by clicking on the buttons in the Dialog Modeler toolbar, running down the left hand margin of the Dialog Modeler window

• from the pop-up context menu, which is displayed by right clicking anywhere in the Dialog Modeler window.

The drawing tools include

Before the dialog is checked out

Zoom in

By a pre-defined factor.

Zoom out

By a pre-defined factor.

Fit to screen

Refresh dialog flow

Page 31

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 31 of 116

When the dialog has been checked out

Arrange steps and connections

Check repository for updated components

If any newer versions are available, SAB gives you the opportunity to update the dialog.

Add component

Opens the Select Component dialog box, which allows you to select a component and then “paste” it in the Dialog Modeler window. See page 39.

Adds a new interrupt component

Allows you to “paste” an Interrupt component in the Dialog Modeler window. See page 66

When a component is selected

Add connection

Note that if the selected component is a dialog step, you must select one of the result bars or this button remains grayed out. See page 41.

Edit selected step

Opens the Edit Step dialog box, which allows you to configure the dialog step. See page 42.

Check if selected step is valid

Checks that the step is connected to the rest of the dialog flow and that all the variables required have been defined. See page 67.

View or edit the master component

Allows you to view detailed general information about the selected component.

Delete selected step.

See page 72.

If you delete a decision point or a rule the associated connections are also deleted.

When you delete a connection, the variables and arrays defined in the earlier in the flow are no longer available to components later in the flow and may need to be re-defined.

Page 32

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 32 of 116

Multiple dialog flow windows

Multiple dialog flow windows may be opened in the Dialog Modeler and can be selected

• from the list displayed under the Windows option in the Menu bar

• from the drop-down list accessed from the blue title bar at the top of the Dialog Modeler

window.

Figure 18. The Dialog Modeler window title bar, showing drop-down list

The dialog flow currently displayed in the Dialog Modeler window can be closed by clicking on the Close button, at the right of the title bar.

Component configuration details

When a component is selected, its configuration details are displayed in the window directly beneath the Browser.

Figure 19. Component configuration details

Page 33

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 33 of 116

The Dialogs tree

Interactions between an automated system and a caller are called dialogs.

The software that performs a dialog is called a dialog flow (or a dialog model or, sometimes, a process). A dialog flow may contain one or more sub-dialog flows (and a sub-dialog flow may have sub flows of its own).

Dialog flows are managed in the Dialogs tree.

Figure 20. The Browser, showing the Dialogs tree

The Dialogs tree provides

• a list or tree of dialog flow folders, in the Browser window

• a Dialog Search engine (see page 73)

• a version control mechanism (see page 37)

• a dialog flow design facility, in the Dialog modeler window (see page 38)

The options presented in the Menu bar and as buttons in the Tool bar are configured dynamically according to the feature you are using.

Page 34

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 34 of 116

Classification

The Classification feature is used to build the file directory structure by grouping and nesting dialog flows; for example, the Figure below shows a dialog flow “Main Dialog” that has been classified under “Direct Debit”, which has itself been classified under “Banking”.

Figure 21. The Dialogs directory structure, showing Classification.

The Classification path is specified when the dialog flow is created or renamed (see page 75).

Adding a new dialog flow

To add a new dialog flow:

1. Click anywhere on the Dialogs tree in the Browser window. Do one of the following:

• Select Dialog Model > Add Dialog Model from the Menu bar

• Click on the Add button and select Add Dialog

• Click on the Add Dialog Model button

• Right click on the Dialogs tree to display the pop-up Dialogs context menu and select Add

Dialog Model.

Page 35

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 35 of 116

SAB displays the first Add New Dialog dialog box.

Figure 22. Add New Dialog 1

2. (Optionally) enter a Classification Path name (see page 34 for an explanation of Classification). You can enter this using one of the following methods:

• one level at a time; for example, enter Banking in the first Classification Path field, Voice

Runner automatically displays another Classification Path field, enter DirectDebit in the second

• in a single string; for example, enter Banking.DirectDebit (separated by a period), in the first

field.

3. Enter a dialog flow Name.

4. Click Finish.

SAB allocates a default version number of 1.0 to a new dialog and automatically adds it to the Tool Environment, where it can be edited.

Page 36

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 36 of 116

It lists the Dialog Model name in the Dialogs tree, and opens a read only version of the new dialog flow, complete with Begin Dialog and End Dialog steps, in the Dialog Modeler window.

Figure 23. The Dialog Modeler window (at right hand side)

Page 37

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 37 of 116

Checking in and out

SAB is a multi-user development environment. To prevent more than one developer simultaneously editing a dialog flow, SAB does not allow you to edit a dialog flow until you have checked it out.

Figure 24. The read-only warning message

While a dialog flow is checked out, no other developer can edit it (though he or she may open another read only version).

The dialog flow should be checked in when you have finished editing.

Checking out

To check out a dialog flow:

1. Click the Check out button in the Tool bar. SAB confirms that the dialog flow has been checked out.

Figure 25. Check out confirmation message

2. Click OK. The background of the Dialog Modeler window turns white and the dialog flow can now be edited.

Page 38

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 38 of 116

Checking in

To check in a dialog flow:

1. Click on the Check in button on the Tool bar. SAB asks if you want to save the dialog flow before you check it in.

Figure 26. Save confirmation message

2. Click Yes to save or No to check in without saving.

Note that you cannot cancel the checking in at this stage. If you want to continue editing, click Yes to save, click OK on the confirmation message (shown in the figure below) then check the dialog flow out again.

3. SAB displays the check in confirmation message.

Figure 27. Check in confirmation message

Building a dialog flow

Dialog flows and sub-flows

Interactions between the caller and the automated system are handled by dialog flows. Depending on the nature of the interaction, a dialog flow may consist of one single-level flow or of a main, top-level, flow with one or more sub-flows. A sub-flow may have sub flows of its own.

Mandatory components

When you open a new Dialog Model, SAB automatically places a Begin Dialog step and an End Dialog step in the Dialog Modeler window. The remainder of the dialog flow is drawn between these

two steps.

A top level dialog flow must end with a Hangup step connected to the End Dialog step.

Page 39

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 39 of 116

Visibility of variables

As you build a dialog flow, SAB automatically makes available to the current component all the variables defined in all the previous components in the path. For this reason, the most efficient way to draw a dialog is to add a component, connect it to the previous component, then edit its details.

Adding a dialog or process component

To add a dialog component to a dialog flow:

1. Click on the Add Component button.

SAB displays the Select Component dialog box.

Figure 28. The Select Component dialog box showing the Search (Dialog) tab selected

2. Select

• the Search (Dialog) tab to find a dialog component

• the Search (Process) tab to find a process component.

3. Do one of the following:

• to find a specific component, type the component name, or partial name, into the Find field

• to find all available components, leave the Find field blank.

A list of components can be found in Appendix A, page 108.

Page 40

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 40 of 116

To direct the search, you can also

• compile a list of exclusion criteria by clicking on Add.

Figure 29. The Select Component dialog box, showing the Add button (center)

SAB displays the Select Categories dialog box.

Figure 30. The Select Categories dialog box, showing a selection of valid categories

Page 41

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 41 of 116

Select the categories to be excluded from the search. Multiple categories may be selected by holding down the CTRL key. Click OK.

A category can be removed from the Exclude List by selecting it then clicking the Remove button.

Avoid unchecking Names, Descriptions, or Keywords, unless you are certain where the name or keyword you have specified appears.

4. Click Search.

SAB displays a list of components in the Results field.

5. Click on the component you require. A brief description of the component appears in the Description field.

6. Click Use.

The cursor changes to a hand holding a component symbol.

7. Click on the canvas of the Dialog Modeler window.

SAB pastes the component onto the canvas. You can reposition it by clicking on it (in the case of a dialog component, click on its title box) dragging it to a new position, and dropping it.

Adding a connection

To add a connection:

1. Click on the first step. In the case of a dialog step, click on the appropriate result bar. The selected step is highlighted in yellow.

Figure 31. First dialog step with result bar selected

2. Click on the Add Connection button.

The cursor changes to a hand holding a connection symbol.

Page 42

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 42 of 116

3. Click on the second step. In the case of a dialog step, click on the title box.

SAB draws the connection, complete with a decision point (the grey diamond shown in the Figure below) that can be used to add a rule to allow branching (see page 64).

Figure 32. A connection between two dialog steps

Setting up or editing a dialog step

To set up or edit a dialog step, double-click on the title box of the component. SAB displays the Edit Step dialog box.

Figure 33. The Edit Step dialog box of the Question component, used to edit dialog steps

Page 43

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 43 of 116

Each dialog step must be configured to perform the specific task required by the dialog flow. The majority of dialog steps handle an interaction with the caller and typically perform the following tasks:

• Play out an initial prompt; for example, Please select one of the following options: Red, Blue or

Yellow (this is configured on the Prompts page of the Edit Step dialog box, see Figure above)

• Recognize the caller’s utterance by matching it against a grammar. The grammar may be

supplied by the component (as in the case of Date, Time and Money recognitions) or it may be supplied by the developer. Some components allow you to specify the grammar as a series of ‘choices’ (typically configured on the Choices and Results page, see the Figure above). Other components allow you to specify an external grammar, written in GSL or JSGF, and stored in a separate file (typically configured on the Grammar Mapping page).

• (Optionally) repeat the result back to the caller and ask him or her to confirm it. (Both the

prompt used and the circumstances under which confirmation is required are configured on the Confirmation page, see the Figure above).

• Handle other behavior by the caller, for example:

o disconfirmation, by playing out a prompt that repeats the initial request to the caller; for

example, Please repeat your selection: red, blue or yellow. (This is configured on the Misrecognition page, see the Figure above)

o silence, by playing out a prompt such as, I’m sorry I didn’t hear that, then repeating the

request to the caller by playing out the Misrecognition prompt. (This is configured on the No Input page, see the Figure above)

o nonsense, by playing out a prompt such as, I’m sorry I didn’t understand that, then

repeating the request to the caller by playing out the Misrecognition prompt. (This is configured on the No Match page).

• For each recognition it performs, the dialog step may also re-configure the speech recognizer.

(This is set up in the Voice Properties dialog box, see page 47).

• A dialog step may also be set to repeat some of these tasks up to a specified maximum

number of times. This is configured on the Settings page.

The following examples illustrate the configuration of a typical component; other components may differ in detail.

Page 44

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 44 of 116

Name and description

To begin configuring a component step:

1. Enter a specific name in the Name field (that describes the purpose of the step).

2. Enter a brief description in the Description field.

Figure 34. The Edit Step dialog box

Prompt

To define the main prompt, select the Prompts tab and do one or more of the following:

• type the text of the prompt into The question to ask field

• if you want to include variable data in the prompt, select the Folder Data tab and drag and

drop the appropriate variable into The question to ask field. Note that the step must be linked to previous steps before SAB can populate the Folder Data field

• if you want to specify an existing .wav file, stored in the SAB database, click on the Prompts

tab, browse to the file and drag and drop it into The question to ask field

• if you want to turn on barge-in click on the Barge-in tab and drag and drop the BARGE-IN

ON command into The question to ask field.

If barge-in is set to ON in Voice Properties, this setting will take precedence over the BARGE-IN ON command and barge-in will be enabled over the entire prompt.

Page 45

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 45 of 116

A prompt can be built up from several of these elements; for example:

Figure 35. Defining a prompt

Dynamic information is formatted (for example, as a date or an amount of money) using variable handlers.

Figure 36. A variable token

Page 46

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 46 of 116

To enable a variable handler:

a. Right-click on the variable token (for example, ~Confirmation Value, in the Figure above). The

SAB displays a Context menu: click on Edit. The SAB displays the Variable Properties dialogue box

Figure 37. The Variable Properties dialogue box

b. Check Substitute matching prompts. This tells SAB to search through the stored prompt files,

select the audio file with the same name as the variable token, and play it out. A tree of prompt folders is displayed in the dialogue box. Select the appropriate folder (this will narrow the search and speed up retrieval of the prompt file at runtime).

c. Check either of the following:

• Match whole phrase only – if SAB cannot find the appropriate file it will leave a “blank” in the

prompt

• Allow mixed TTS/prompts – if SAB cannot find the appropriate file it will use TTS to supply

the missing word or phrase.

Page 47

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 47 of 116

Select the appropriate variable handler. The options include:

none

Base64binary

Use when playing out a prompt file recorded using a RecordDialogStep. Messages recorded by RecordDialogStep (WAV files) are encoded in base 64 so that they can be stored as XML. The Base64binary handler restores the file to its original format so that it can be played to the caller outside the RecordDialogStep.

Default

Digit Sequence TTS

Reads a string of digits as a sequence of digits; for example, three, nine, one not three hundred and ninety one.

Character Prompt

Reads a single character. Typically used for constructing phonetic alphabet prompts; for example, Was that A for Alpha?

Date

Use with any of the Date dialog components. Reads a numerical variable as a date.

Money

Use with any of the Money dialog components. Reads a numerical variable as an amount of money.

Voice Properties

To set the “Voice Properties” (the recognition parameters for this recognition), click Show Details. SAB displays the Voice Properties dialog box.

Figure 38. The Voice Properties dialog box

Page 48

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 48 of 116

SAB allows you to configure the following recognition parameters:

Parameter

Description

Timeout

The Timeout parameter sets the length of silence that must elapse before the recognizer returns a value of No Input.

Barge-in

The Barge-in parameter specifies the barge-in setting to be used on this recognition. Options are:

• ON

• OFF

• DEFAULT

If Barge-in is set to ON here, it is enabled on the entire prompt; if it is set to OFF here, it can be turned on at a specific point in the prompt using the BARGE-IN ON command. See page 44.

Complete Timeout

The Complete Timeout parameter sets the length of silence that must follow the caller’s utterance before the recognizer assumes that the utterance has ended, and finalizes a result. The result may be a successful recognition or a nomatch.

If Complete Timeout is set too low, the recognizer may timeout before the caller has finished speaking (for example, in a pause between phrases).

If Complete Timeout is set too high, recognition time will be unnecessarily increased.

Note that the recognizer times out after Complete Timeout when it has a complete match to an active grammar; it uses Incomplete Timeout when the speech is an incomplete match to an active grammar. Complete Timeout should always, therefore, be set to a slightly lower value than Incomplete Timeout.

Incomplete Timeout

The Incomplete Timeout parameter sets the length of silence that must follow the caller’s utterance before the recognizer finalizes a result. This value is used when the caller’s utterance (prior to the silence) is an incomplete match to all active grammars. The partial result is classified as a no-match.

Incomplete Timeout is also used when the caller’s utterance (prior to the silence) is a complete match to an active grammar, but it would be possible for him or her to speak further and still match the grammar.

If Incomplete Timeout is set too low, the recognizer may timeout before the caller has finished speaking (for example, in a pause between phrases).

If Incomplete Timeout is set too high, recognition time will be unnecessarily increased.

The Incomplete Timeout should be set to a slightly higher value than Complete Timeout, to allow callers to pause mid-utterance.

Page 49

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 49 of 116

Parameter

Description

Confidence Level

Confidence is a floating point value that indicates the perceived accuracy of the recognition result.

The Confidence Level parameter sets the accuracy threshold: a recognition result with a confidence value above the threshold is considered successful; a recognition result with a confidence value below the threshold is considered a no-match.

Sensitivity

The Sensitivity parameter configures the recognizer to handle the expected signal level. A value of 100 is highly sensitive (suitable for quiet input); a value of 0 is least sensitive.

Note that the setting required to handle a given signal level will depend on the recognizer being used.

Speed VS Accuracy

The Speed VS Accuracy parameter controls the trade-off between recognition speed and recognition accuracy. A value of 0 specifies the fastest recognition; a value of 100 specifies highest accuracy.

Universal

Some platforms provide optional universal command grammars, to recognize utterances such as "Help", "Cancel", or "Exit", which are always active.

The Universal parameter is used to specify which universals are to be handled in this recognition. The options include

• none

• all

• help exit

• help cancel

• help cancel exit

• help cancel repeat

• help cancel exit repeat

These are selected from a drop-down list.

Note that the availability of this feature is dependent-dependent.

Input Mode

The Input Mode parameter specifies which mode or modes the caller can use to respond for this recognition. The options are

• dtmf voice (either)

• dtmf

• voice

These are selected from a drop-down list.

Note that this parameter does not activate or de-activate speech recognition grammars.

Page 50

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 50 of 116

Parameter

Description

Fetch Audio

The Fetch Audio parameter specifies the location of the audio clip to be played to cover the delay while waiting for a document to be fetched. The fetching of the audio clip is governed by the values of Audio Fetch Hint and Fetch Timeout.

The playing of the audio clip is governed by the values of Fetch Audio Delay and Fetch Audio Minimum.

Fetch Timeout

The Fetch Timeout parameter sets the maximum length of time for fetches.

Fetch Audio Delay

When a fetch delay is short, it may be better to allow a few seconds of silence rather than play a very short snatch of audio.

The Fetch Audio Delay parameter sets the time interval that must elapse at the start of a fetch delay before the audio clip is played. See Fetch Audio Minimum.

The default interval is platform-dependent.

Fetch Audio Minimum

The Fetch Audio Minimum parameter specifies the minimum length of time an audio clip is played, once started, even if the fetch result arrives in the meantime. See Fetch Audio Delay.

The default is platform-dependent.

Audio Fetch Hint

The Audio Fetch Hint parameter specifies whether the platform can attempt to optimize dialog interpretation by pre-fetching the audio. The options are

• safe - the platform fetches the audio when it is needed

• prefetch - the platform is permitted, but not required, to pre-fetch the

audio

• default

These are selected from a drop-down list.

Once you have set the appropriate voice properties, click OK.

Page 51

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 51 of 116

Confirmation

If the caller’s answer requires confirmation select the Confirmation tab and do the following:

• define the confirmation prompt as described in the previous step. A confirmation prompt

typically has the following format:

Typed

Variable dragged and dropped from

a list of previously defined variables

Typed

Command dragged

and dropped

Typed

Did you

say

<caller_data>

<BARGE-IN_ON>

Please say

yes or no.

Figure 39. A confirmation prompt

• if you want the step to confirm the data it has captured in every case, check the Explicit

confirmation box.

Figure 40. Setting up confirmation handling 1

Page 52

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 52 of 116

• if you want the step to confirm the data it has captured only if the recognition result has a

confidence value below a certain threshold, leave the Explicit confirmation box unchecked and set the slider controls to the thresholds you require

Figure 41. Setting up confirmation handling 2

When using a Nuance recognizer: to set a confidence value of approximately zero, position the pointer slightly above zero. If you set it to zero SAM will use the Global Setting of 50.

o the Acceptance Confidence Threshold sets a confidence limit on the recognition result.

If the result has a confidence value above this threshold, it is accepted without confirmation. If it has a confidence value below this threshold it is played back to the caller who is asked to confirm that it is correct

o the Acceptance Confidence Margin specifies the margin by which the result with the

highest confidence level must beat the remaining results to be considered correct

o the Confirmation Acceptance Confidence Threshold sets a confidence limit on the

recognition result of the confirmation dialog.

• Set the Voice Properties (Confirmation Voice Dialog Properties), see page 47.

• A caller who responds to the confirmation question with a no-match utterance (nonsense) or

with silence, is handled by setting Confirmation No Match and Confirmation No Input, respectively.

Page 53

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 53 of 116

To set either of these functions, click the appropriate Show Details button.

Figure 42. Setting up confirmation handling 3

SAB opens a dialog box; for example, the Confirmation No Match dialog box is shown in the Figure below.

• Define the prompt wording.

• Configure the recognition parameters (Voice Properties). Different sets of voice properties may

be configured for each recognition.

A step may be configured (via the Settings page, shown in the Figure above) to re-prompt the caller once, or more than once, and each prompt may be worded differently. The dialog box defaults to the Level 1 prompt; to add another prompt, click Add.

Figure 43. The Confirmation No Match dialog box

When you have defined the prompt or prompts, click Close.

Page 54

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 54 of 116

Note that confirmation is not usually required when the main prompt is a yes-no question, but is often necessary if the caller has provided accurate data, such as a postcode or an amount of money.

Speech recognition

Components that are designed to recognize a standard item of data, such as a date, a time or a weight, have built-in grammars that will recognize any valid response. For these components it is typically only necessary to specify a result variable. This is done on the Results page.

Figure 44. The Edit Step dialog box of the Weight component, showing the Results page

Page 55

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 55 of 116

Other components are designed to capture information specified by the developer.

Choices

For example, for the Question Dialog Step you must define a set of valid caller responses, with synonyms (alternative words) if appropriate, and SAB uses these to generate a recognition grammar.

These recognitions are typically configured on the Choices and Results page.

Figure 45. The Choices and Result page of the Question component Edit Step dialog box

To add a valid response:

1. Click Add Value. SAB displays Add Value dialog box.

Figure 46. The Add Value dialog box

2. Enter the word to be recognized in the Return Value field.

3. (Optionally) select a DTMF equivalent from the drop-down list in the DTMF Code field.

Page 56

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 56 of 116

4. To (optionally) enter a synonym, click the Alternatives : Add button. SAB displays the first page of the Attach Values wizard.

Figure 47. Attach Values 1

5. Click Next. SAB displays the second page.

Figure 48. Attach Values 2

6. Enter the synonym in the Substitute Word field. Click Finish. SAB re-displays the Add Value dialog box, with the alternative word shown in the Alternatives field.

Page 57

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 57 of 116

7. Repeat from Step 4 to define further synonyms.

Once the word or phrase, its DTMF equivalent, and its synonyms are defined, click OK.

Figure 49. The Add Value dialog box, showing a valid response

SAB displays the option in the Valid Answers field of the Edit Step dialog box.

Figure 50. The Edit Step dialog box of the Question component, showing a valid response

defined

To define another valid response, repeat from Step 1 above.

Select or create the result variable.

External grammars

Some components, such as the SpecifiedGrammarQuestionDialogStep, allow you to specify more complex grammars, which are defined using GSL (for a Nuance speech recognizer) or JSGF (for an IBM speech recognizer) and saved in a file labeled <grammar_name>.grammar or <grammar_name>.jsgf respectively.

Page 58

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 58 of 116

To specify the grammar:

1. Open the Edit Step dialog box and select the Grammar Mapping page.

Figure 51. The Grammar Mapping page

2. Enter the grammar rule name in the Mapping Grammar Rulename field.

3. Enter the grammar filename in the Mapping Grammar FileName field. It is not necessary to include the file extension.

4. Specify how the recognition results are handled by defining the Slot Names as follows:

a) Click on Add Value. SAB displays the Add Value dialog box.

Figure 52. The Add Value dialog box

b) Enter the first Slot Name associated with the grammar rule.

c) Enter the variable into which the value associated with the slot (the semantic value of the

recognition result) is to be written.

d) Click OK.

e) Repeat from step a) until all the Slot Names associated with the grammar have been defined.

Page 59

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 59 of 116

No Input

If the caller remains silent, the dialog step plays the prompt that has been defined on the No Input page. For an explanation of prompt levels, see page 53.

Figure 53. The No Input page, showing a typical No Input prompt.

No Match

If the caller says something that does not match any of the choices defined on the Choices and Results page, the dialog step plays the prompt that has been defined on the No Match page. For an explanation of prompt levels, see page 53.

Figure 54. The No Match page, showing a typical No Match prompt.

Page 60

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 60 of 116

Misrecognition

If the caller disconfirms the recognition result, the dialog step plays the prompt that has been defined on the Misrecognition page. For an explanation of prompt levels, see page 53.

Figure 55. The Misrecognition page, showing a typical Misrecognition prompt.

Settings

The Settings page is used to configure the general behavior of the dialog step. Its contents are component specific. For example, the Settings page of the Question dialog step allows you to set:

• Maximum reprompts allowed for confirmation

• Maximum reprompt

• Maximum disambiguation prompts allowed.

The Settings page of the Call Redirect dialog step allows you to:

• set the target address

• enable or disable tromboning

And the Settings page of the Number Sequence dialog step allows you to set:

• Minimum Length

• Maximum Length

• Validation

• Maximum reprompts allowed for confirmation

• Maximum reprompts

• Maximum disambiguation prompts allowed

Page 61

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 61 of 116

Setting up or editing a process component

Process steps handle data. SAB processes include:

• addition, subtraction, multiplication, division, total, exponentiation

• check lookup table

• capitalize, change case

• set variable, copy variable, insert new variable

• report call end, report call information

• send event code

and so on. The data being manipulated may be static, or derived from a recognition result, or retrieved from a database.

To set up a process component:

1. Double click on the process step icon. SAB displays the Edit Step dialog box.

2. The majority of Process components are configured via a single page, called Default. The data that must be entered is component-specific. As you build a dialog flow, SAB automatically makes available to the current component all the variables defined in all the previous components in the path. These are selected from drop-down lists.

Figure 56. An example of a Process component Edit Step dialog box

3. A step trigger may be set up as described below.

Page 62

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 62 of 116

Setting up a trigger

There are five types of trigger:

• Exception - evaluated only when an exception occurs within a dialog flow

• Before step validated - legacy feature - DO NOT USE

• Before step runs - evaluated before the step

• After step runs - evaluated after the step

• Unhandled control - evaluated if the path required is not defined in the dialog flow.

Triggers may be configured

• as part of a flow, in the Trigger Modeler window. These triggers are used by default when any

triggering event occurs in that flow, if the step in which the event occurs does not specifically

handle triggers

• as part of a step, in the Edit Step dialog box. Step-based triggers are specific to the particular

instance of the component and take priority over any global triggers.

Handling exceptions

If an exception occurs in a flow, and no Exception trigger has been defined, the flow will fail.

There are three ways handle an Exception trigger.

• try the step again

• skip over the offending step

• elegantly end the call

To configure the flow either to try the step again or to skip over the step, it is necessary to include an

ExceptionFlowControlStep in the trigger dialog flow (see page 63). However, if the ExceptionFlowControlStep is not defined, the flow will throw and unhandled exception and the caller

will hear the default error message defined by the Voice Stack that is being used.

To end the call elegantly, you must build a trigger dialog flow that hangs up the call and add a TerminateStep at the end of the flow. This will instruct the runtime that the flow must end immediately and the flow will not throw an unhandled exception.

Flow trigger

To set up a flow trigger:

1. Select the Trigger Modeler window.

2. Select the type of trigger you want to create (see list above).

3. Check Create. SAB opens a dialog window.

4. Build the alternative dialog flow. If the trigger being defined is an Exception trigger, ensure that the dialog flow contains an ExceptionFlowControlStep (see below).

Page 63

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 63 of 116

Step trigger

To set up a step trigger:

1. Open the Edit Step dialog box.

2. Select the Edit Step trigger page.

3. Create the trigger as described above.

ExceptionFlowControlStep

An exception trigger dialog flow must always contain at least one ExceptionFlowControlStep, which must be in a path that will be executed whenever the trigger dialog flow is run.

The ExceptionFlowControlStep is used to specify how the runtime environment should return from the exception trigger.

Figure 57. The ExceptionFlowControlStep

To configure the step:

1. Select the appropriate Runnable Step Assignment id from the drop down list. This indicates the step that has triggered the exception.

2. Select the instruction required. The ExceptionFlowControlStep provides two options:

• retry - execute the step that generated the exception again

• continue - execute the step after the step that generated the exception.

Page 64

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 64 of 116

Adding a rule

To define a rule:

1. The order in which you draw the connections between the source step and the steps to be executed is used by SAB to determine how to interpret the rule.

First, add a connection between the source step and the step that must be executed if the rule yields a TRUE result; secondly, add a connection between the source step and the step that must be executed if the rule yields a FALSE result.

2. Double click on the decision point on the TRUE path. SAB opens the Rule Editor dialog box.

Figure 58. The Rule Editor dialog box

3. Enter a brief description. This will be displayed next to the rule in the Dialog Modeler window.

4. Click on the Select Component button. SAB displays the Select Component dialog box.

Page 65

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 65 of 116

5. Select the component you require, using the search facility (see page 39).

Figure 59. The Select Component dialog box

6. Click Use. SAB displays the undefined rule in the Rule Editor dialog box.

Figure 60. The undefined rule

Page 66

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 66 of 116

7. Double click on the rule icon (“less than or equals to” in the Figure above). SAB displays the Edit Rule dialog box.

Figure 61. The Edit Rule dialog box

8. The data you will need to enter is rule-specific. As you build a dialog flow, SAB automatically makes available to the current component all the variables defined in all the previous components in the path. These are selected from drop-down lists.

9. Click OK.

Adding an Interrupt Step

An Interrupt Step step handles inputs from the caller (such as “Help” or "Cancel") that may happen at any point in the dialog.

An Interrupt Step can be added at any point in the dialog flow. It must have no entrance point, but should lead to a (typically short) dialog flow designed to handle the caller’s command.

Once an interrupt step has been added to a flow, the specified phrase will interrupt any step in the same flow that performs a recognition (that can detect what the caller has said).

If required, when the alternate flow has completed, you can route the caller back to the step that was interrupted, using a ReturnFromInterrupt process step.

To add an Interrupt step:

• Click on the Interrupt Step icon on the vertical toolbar, then drop the component in the Dialog

Modeler window, as normal.

To configure an Interrupt step:

1. Double click on the component. SAB displays the Edit Step dialog box.

2. Click Add Value. SAB displays the Add Value dialog box.

Figure 62. The Add Value dialog box

3. Enter the Interrupt trigger (such as help or cancel).

Page 67

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 67 of 116

4. Click OK. The Interrupt trigger is displayed in the Interrupt Trigger Conditions field. To define another trigger, repeat the process.

5. Click OK.

Validating a step

The Validate Step tool checks the selected step has been correctly configured.

To validate a step:

1. Click on the step.

The step is highlighted in yellow.

2. Do one of the following:

• right-click on the component to display the pop-up Context menu and select Validate Step

• click on the Validate Step button on the Dialog Modeler Tool bar.

If the step is valid, SAB displays the result in a pop-up message:

Figure 63. Step valid message

If the step is invalid, SAB displays a pop-up error message, which lists the errors it has detected and provides links to the Java stack trace.

Figure 64. Step invalid error message

Page 68

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 68 of 116

Sub-dialogs

A dialog flow can be inserted into another dialog flow as a sub-dialog.

Any data to be passed into the sub-dialog from the calling dialog flow must be specified in the Required Data window, and then assigned in the Sub Dialog step; any data to be passed out of the sub-dialog to the calling dialog must be specified in the Output Data window, and then assigned in the Sub Dialog step.

The Required Data window

Data to be passed from a dialog flow to a sub-dialog flow must be specified in the Required Data window.

To define an input variable:

1. With the sub-dialog flow open in the Dialog Modeler window, click on the Required Data tab. SAB displays the Required Data window.

2. Click Add. Voice Runner displays the Add Data Type wizard.

3. Select the data type required.

4. Click Next

5. Enter a variable name and set the maximum and minimum occurrences (set values of 1 for a variable or set the appropriate values for an array).

6. Click Finish.

The variable is listed in the Required Data window.

Variable names must not

• start with a number or punctuation

• start with “xml” or “XML”

• contain spaces.

See Appendix B, page 113 for a complete list of data types.

The Output Data window

Data to be passed from a sub-dialog flow back to the calling dialog flow must be specified in the Output Data window, and then declared in the Sub Dialog step.

To define an output variable:

1. With the sub-dialog flow open in the Dialog Modeler window, click on the Output Data tab.

2. Click Add Assignment. SAB displays the New Datatype wizard, which lists the data items associated with the current dialog flow.

3. Select the data item required.

4. Click Next.

5. Enter a variable name.

6. Click Finish.

Page 69

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 69 of 116

The variable is listed in the Output Data window.

Adding a sub-dialog to a dialog flow

To add a sub-dialog to a dialog flow:

1. Click on the Add Component button.

SAB displays the Select Component dialog box.

2. Select the Dialog Models tab.

Figure 65. The Select Component dialog box showing the Dialog Models tab selected

3. Browse to the dialog flow you require and highlight it.

4. Click Use.

The cursor changes to a hand holding a component symbol.

5. Click on the canvas of the Dialog Modeler window.

SAB pastes the dialog flow onto the canvas as a single component. You can reposition it by clicking on its title box, dragging it to a new position, and dropping it.

Figure 66. A sub-dialog component

Page 70

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 70 of 116

To assign the input variables:

1. Double click the Sub-Dialog step. SAB displays the Step Editor, which lists, in the Sub-Process input column of the Input tab, the input variables that you have already specified in the Required Data window.

Figure 67. The Sub Dialog step

2. Select the variable you want to assign.

3. Click Edit. SAB lists the variables visible to the sub-dialog component.

4. Select the variable you require and click OK.

To assign an output variable:

5. Select the Output tab.

Page 71

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 71 of 116

6. Click Add Assignment. SAB lists the output variables you have already defined in the Output Data window.

Figure 68. The Output Data window

7. Select the variable you require and click Next.

8. Select Add value to folder (add variable) or Replace value in folder (overwrite variable). Click

Next.

Figure 69. The Output Mapping dialog box

Page 72

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 72 of 116

9. Do one of the following:

• if you have opted to add a variable, enter a variable name

• if you have opted to overwrite a variable, select the variable you want to overwrite.

10. Click Finish.

Deleting a dialog flow

To delete a dialog flow:

• Do one of the following:

o right-click on the dialog flow icon (not the version number) to display the pop-up

Context menu and select Delete Dialog Model

o select the dialog flow icon (not the version number) and select Dialog Model > Deletes

the Selected Dialog Model Version

SAB asks you to confirm that you want to delete the dialog flow. Click Yes. The dialog flow is deleted.

Deleting components and connections

A component must be physically connected to another component to see its variables. If you delete a component or a connection, and break the flow of the dialog, subsequent components will lose all reference to variables from components earlier in the flow (the fields will be blank). To replace the variables it is necessary to restore a connection, then edit each component and re-select the variable names. To avoid this, always draw an alternative connection (possibly temporary), before deleting a component or a connection.

Deleting a dialog or process component

To delete a dialog or process component:

1. Click on the component to highlight it.

2. Do one of the following:

• Right-click on the component to display the pop-up Context menu then select Delete Step

• Click on the Delete button on the Dialog Modeler Tool bar

Deleting connections

To delete a connection:

1. Click on the rule, or on the undefined decision point, that lies on the connection you want to delete.

2. Do one of the following:

• Right-click on the component to display the pop-up Context menu then select Delete Decision

• Click on the Delete button on the Dialog Modeler Tool bar

Page 73

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 73 of 116

Deleting rules

If you use the method described in the previous section to delete a rule, SAB will also delete the connection it lies on.

To delete a rule:

1. Double click on the rule. Voice Runner opens the Rule Editor dialog box.

2. Right click on the rule icon.

3. Select Delete Rule.

4. Click OK.

Finding a stored dialog

The Dialogs tree

To find a stored dialog: expand the Dialogs tree and select the dialog you require.

• Click on the dialog folder to display the version index.

• Click on the dialog version number to open a read-only version of the dialog in the Dialog

Modeler window.

The version index consists of two tabbed worksheets:

• Dialog Information - which lists the available versions of your selected dialog flow, and

indicates the environment in which they operate.

• Authorized Members - which lists the dialog flow’s owners and authorized members (users

who are permitted to edit the dialog flow).

The search engine

SAB provides a search engine that can be used to search for the dialog flow you require. To search:

1. Click on a parent node in the Dialogs tree. Voice Runner displays its Dialog Search engine in the Dialog modeler window.

2. Type the name, or partial name, of the dialog flow you want to find into the Name field. If you have used a partial name, ensure that the Partial Match box is also checked.

3. Select the environment you want to search:

• Any environment - searches all environments; this is the default

• Tool - searches the Tool environment. This contains dialog flows that are under construction

(and related prompts)

• Production - searches the Production environment. This contains dialog flows that are

deployed (and related prompts)

• Decommission - searches the Decommission environment. This consists of dialog flows that

have been taken out of deployment (and related prompts).

4. Click Search. SAB displays a list of all matching dialog flows in the Dialogs modeler window.

5. Click Show Details to display all available versions of the dialog flow.

Page 74

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 74 of 116

6. Click on the version number to open a read-only copy of the dialog flow.

Before you can edit a dialog flow, you must check it out, see page 37.

Copying a dialog flow

Copy with the same version number

To copy a dialog flow with the same version number:

1. Right click on the dialog flow icon (not the version number) to display the pop-up Context menu.

2. Select Copy dialog model and a selected version. SAB displays the Copy Dialog Model dialog box.

Figure 70. The Copy Dialog Model dialog box

3. By default, SAB assigns the name “Copy of...” to the new dialog; this may be changed. If more than one version of the dialog is available, select the version number you require.

4. Click Finish. SAB displays the new dialog flow in the Dialogs tree. Note that files are displayed in alphabetical order, so “Copy of...” will be displayed near the top of the tree.

Copy with an incremented version number

Before a dialog flow can be copied with an incremented version number it must be migrated to the Production Environment, see page 76. The new dialog flow is created in the tool environment.

To copy a dialog flow with an incremented version number:

1. Do one of the following:

• right-click on the dialog flow icon (not the version number) to display the pop-up Context menu

and select Increments the version of the selected dialog version

• open the dialog flow and click on the Increments the version of the selected dialog version

button or select Dialog Model > Increments the version of the selected dialog version.

Page 75

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 75 of 116

2. SAB displays a list of dialog flows.

Figure 71. The Increments the version of the selected dialog version button dialog box

3. Select the dialog flow you want to copy. Click Next.

4. Select the version number you want to copy. Click Next.

5. Click Finish. The new dialog flow is created in the tool environment. SAB displays it in the Dialogs tree and automatically opens an editable version in the Dialog Modeler window.

Renaming a dialog flow

If the dialog flow you are renaming has a classification and you want to retain it, you must include the classification path as part of the new name.

To rename a dialog flow:

1. Right-click on the dialog flow icon (not the version number) to display the pop-up Context menu and select Rename Dialog Model. SAB displays an Input dialog box.

Figure 72. Input dialog box

2. Enter the new name. If the dialog flow you are renaming has a classification and you want to retain it, include the classification path as part of the new name; for example, Banking.DirectDebit.MainFlow specifies a flow name MainFlow, located in the DirectDebit folder, which is located in the Banking folder.

3. Click OK.

Page 76

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 76 of 116

Migrating a dialog flow

To migrate a dialog flow from the Tool environment to the Production environment:

1. Open the dialog flow.

2. Check the dialog flow out (see page 37).

3. Click on the Migrate button. SAB will ask you to confirm that you want to migrate the dialog flow.

4. Click Yes.

5. SAB will check if the dialog flow is valid.

• If the dialog flow is not valid, it displays a list of errors.

• If the dialog flow is valid, it migrates the dialog and displays a read only version in the Dialog

Modeler window.

There is now a limited set of operations you can perform on the dialog flow:

• Save

• Export

• Copy with incremented version number (see page 74)

• Check in

• Decommission

• Generate war file for deployment

Decommission

To decommission a Production Environment dialog flow:

Open the dialog flow.

1. Click on the Decommission button. SAB will ask you to confirm that you want to decommission the dialog flow.

2. Click Yes.

Page 77

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 77 of 116

Saving a dialog flow

By default, dialog flows are saved in the SAB database. However, they can also be exported to an external file (see page 77).

Saving a dialog flow to the SAB database

To save an open dialog flow:

• Do one of the following:

o Select Dialog Model > Save

o Click on the Saves Current Dialog Model Version

Exporting one or more dialog flows to file

To upgrade the SAB Configurator it is necessary first to uninstall the existing version. When SAB is uninstalled all dialog flows and related resources stored within the SAB database are also deleted. It is therefore essential to export dialog flows to file before uninstalling SAB. The dialog flows can be re-imported after the Configurator has been upgraded.

Multiple dialog flows

To export multiple dialog flows:

1. Expand the Dialogs tree.

2. Do one of the following:

• Click on the parent dialog, then select Dialog Model > Exports Multiple Selected Dialog

Models to a File from the Menu bar

• Click on the parent dialog, then click the Exports Multiple Selected Dialog Models to a File

button

• Right click on the parent dialog to display the pop-up Dialogs context menu and select

Exports Multiple Selected Dialog Models to a File.

Page 78

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 78 of 116

SAB displays the Export Dialog Model dialog box, which lists all the dialogs available.

Figure 73. Export Dialog Model dialog box 1

3. Check the Dialog flows you want to export.

4. Click Next.

SAB lists the available models of the dialog flows you have selected.

Figure 74. Export Dialog Model dialog box 2

5. Check the versions you want to export

6. Click Finish. SAB displays the Save dialog box.

7. Browse to the appropriate location and click Save.

Page 79

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 79 of 116

Single dialog flow

To export a single dialog flow:

1. Expand the Dialogs tree.

2. Do one of the following:

• Click on the dialog flow, then select Dialog Model > Export the Dialog Model to a File from

the Menu bar

• Click on the dialog flow, then click the Export the Dialog Model to a File button

• Right click on the dialog flow to display the pop-up Dialogs context menu and select Export

the Dialog Model to a File

SAB displays the Export Dialog Model dialog box, which lists all available versions of the selected dialog flow.

Figure 75. Export Dialog Model dialog box 1

3. Check the versions you want to export.

4. Click Finish. SAB displays the Save dialog box.

5. Browse to the appropriate location and click Save.

Importing a dialog flow from file

To import a dialog flow from file:

1. Click on the Dialogs tree in the Browser window. Do one of the following:

• Select Dialog Model > Imports a Dialog Model from File from the Menu bar

• Click on the Imports a Dialog Model from File button

• Right click on the Dialogs tree to display the pop-up Dialogs context menu and select Imports

a Dialog Model from File

2. Browse to the location of the .vra file, select it and click Open.

If the .vra file contains out of date components, SAB displays a warning and may ask you to resolve any step conflicts that arise (see page on page 80).

Page 80

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 80 of 116

3. (Optionally) enter an new dialog flow name.

Figure 76. Change name of imported dialog flow

4. Click Finish.

Problems with importing

If the .vra file contains out of date components, SAB displays the following warning.

Figure 77. Out of date components warning

Page 81

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 81 of 116

5. Click Next.

SAB displays a table of missing component versions and their possible replacements.

Figure 78. Component replacements

6. Click Finish.

Step conflict

If a component cannot automatically be upgraded to a new equivalent, SAB displays Step Conflict information.

Figure 79. Step conflict dialog box

To resolve the step conflict:

1. Either select the appropriate new component from the drop down list or select Leave unconnected to delete the path.

2. Click OK.

Page 82

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 82 of 116

The Prompts tree

Prompts are questions or messages that are played out to the caller as part of an interaction (dialog). Prompts may be created as audio (.wav) files or as text, which is converted to electronic speech using a Text-To-Speech (TTS) Synthesizer.

Prompts are managed using the Prompts resource.

The Prompts resource provides

• a list or tree of prompt files, displayed in the Browser window

• a Prompt Search engine

• a simple prompt recording and editing facility, displayed in the Dialog Modeler window.

The options presented in the Menu bar and as buttons in the Tool bar are configured dynamically according to the feature you are using. For a complete list of the Prompts menu options, see page on page 20; for a complete list of Prompts buttons see page on page 23.

Adding a new recorded prompt

To add a new recorded prompt:

1. Click on Prompts in the Browser window. Do one of the following:

• Select Prompt > Add Prompt from the Menu bar

• Click on the Add button and select Add Prompt

• Click on the Add Prompt button

• Right click on the Prompts tree to display the pop-up Prompts context menu and select Add

Prompt

SAB displays the first Add New Prompt dialog box.

Figure 80. Add New Prompt 1

Page 83

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 83 of 116

2. (Optionally) enter a classification path name (see page 34 for an explanation of classification).

3. Enter a prompt name.

4. Click Next.

Figure 81. Add New Prompt 2

5. Specify the source of the new prompt - an existing file or a new recording.

6. Click Next.

Page 84

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 84 of 116

Specifying an existing file

If you have opted to use an existing file, SAB opens a file selection window.

Figure 82. Add New Prompt file selector

1. Select the .wav file required.

2. Click Next. Save the prompt (see page 86).

Recording a new prompt

If you have opted to record a new prompt, SAB displays the recording tool shown below.

Figure 83. Add New Prompt 3: recording tool

Page 85

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 85 of 116

The following options are provided by the Prompt editor.

Load Prompt from File

Record Prompt

Play Prompt

Pause Prompt

Stop Prompt

Save Prompt to File

Remove selected audio

Remove all but selected audio

Clear

To record a prompt:

1. Connect a headset and microphone to the sound card on your computer.

2. Click the Record button and say the prompt.

3. Click the Stop button to stop recording.

To hear the recording, click the Play button.

To re-record the prompt, follow the recording procedure described above.

Page 86

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 86 of 116

Saving the prompt

To save the prompt, when you have either selected the correct .wav file or are satisfied with the recording:

1. Click Next. SAB displays the final dialog box:

Figure 84. Add New Prompt 4

2. Enter the exact text of the prompt you have recorded into the top field. This text will be used under the following circumstances:

• it is played out using TTS if the .wav file is not available (to prevent the dialog coming to a

halt)

• it is used to select the correct .wav file when you are concatenating prompts to generate a

longer prompt

so it is important to enter the wording correctly.

3. Select the prompt language from the drop down list. This information is used to select

• the correct version of the prompt file (you can design an application to offer multiple languages

by creating identically named sets of prompts in different languages)

• the correct TTS engine

4. Select the persona from the drop down list (see page 90 for an explanation of persona). This information is used to select

• the correct version of the prompt file (you can design an application to offer multiple personae

by creating identically named sets of prompts in different persona types)

• the correct TTS engine

5. Click Finish.

The new prompt file appears, at the appropriate position in the Prompts tree, in the Browser window.

Page 87

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 87 of 116

Searching for a prompt

SAB provides a search engine that can be used to search for the prompt you require. To search:

1. Click on the Prompts tree. Voice Runner displays its Prompt Search engine in the Dialog Modeler window.

2. Type the name, or partial name, of the prompt you want to find into the Name field. If you have used a partial name, ensure that the Partial Match box is also checked.

3. Select the environment you want to search:

• Any environment - searches all environments; this is the default

• Tool - searches the Tool environment. This contains dialog flows that are under construction

(and related prompts)

• Production - searches the Production environment. This contains dialog flows that are

deployed (and related prompts)

• Decommission - searches the Decommission environment. This consists of dialog flows that

have been taken out of deployment (and related prompts).

4. Click Search. SAB displays a list of all matching prompts in the Dialogs modeler window.

5. Click Show Details to display all available versions of the prompt.

Viewing a recorded prompt

Prompt files in the Tool environment can be edited. To open a .wav file for editing:

Either select the prompt in the Prompts tree or use the Prompt Search facility to locate the file (see above). SAB lists the available versions of the prompt in the Dialog Modeler window; you can click

Play to hear the prompt, and on Stop to stop playout, if necessary

Figure 85. Prompt list

Click on the prompt version number to open the prompt editor.

Figure 86. Prompt editing

Use the drop down arrow (in the center of the title bar) to list the dialog version associated with this prompt.

Page 88

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 88 of 116

Editing a prompt

A recorded prompt can be edited; for example a long silence at the start of the prompt can be removed. To edit a prompt:

1. Using the mouse, highlight the section of the prompt to be removed. (Place the cursor at the start of the section, press and hold the left mouse button, drag the cursor to the end of the section, then release the mouse button).

Figure 87. Editing a prompt

2. Click the Scissors button to delete the selected area.

Alternatively, you can delete everything but the selected area by clicking on the Curly brackets button.

3. When you select another resource, SAB asks whether you want to save your changes.

Figure 88. Save prompt dialog box

4. Click Yes.

Page 89

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 89 of 116

Exporting and importing prompt files

All SAB prompts are stored by default in the SAB database.

Saving a prompt to file

To save a prompt to file:

1. Open the prompt .wav file

2. Click on the Disk icon. SAB displays a Save dialog box.

3. Browse to the file location and enter the appropriate filename.

4. Click Save.

Saving a prompt directory as a .zip file

To save a prompt directory as a .zip file:

1. Do one of the following:

• click on the directory name in the Prompts tree, then select Prompt > Create Wave Extract

• right click on the directory name in the Prompts tree to display the pop-up menu, then select

Create Wave Extract.

2. Browse to the file location and enter the appropriate filename. It is not necessary to include the .zip extension.

3. Click Save.

Importing prompts from a .zip file

To import prompts from a .zip file:

1. Do one of the following:

• click on the name of the directory (in which you want to store the imported prompts) in the

Prompts tree, then select Prompt > Import Wave Zip

• right click on the name of the directory in the Prompts tree to display the pop-up menu, then

select Import Wave Zip.

2. Browse to the location of the .zip file and select it.

3. Click Open.

The new prompts directory is displayed in the Prompts tree.

Refresh

If the imported prompts do not automatically appear in the Prompts tree:

• Right click on the parent directory to display the pop-up menu and select Refresh Tree From

Here.

This option, placed strategically at various nodes of the directory tree, refreshes all the child nodes of the selected node.

Page 90

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 90 of 116

Persona

What is persona?

A persona is the “personality” one person assigns to another person (real or virtual) based on how he or she looks or sounds, or seems to behave. In the case of an automated telephone service, persona is created by details such as

• the voice type – sex, age, accent

• the language used – formal, informal

• the speed at which it “talks”

• more subtle qualities – such as perceived friendliness, authority or patience.

When a person interacts with another person, both parties assume well-defined roles that allow them to interpret the other’s intentions, anticipate what he or she will do next, and react accordingly. One way to create an automated telephone service that is easy to use is to give it a consistent persona that is appropriate to the task it is performing.

There are several advantages to doing this:

• if you do not, callers will by default. Research has shown that, when using an automated

service, people automatically respond to the system as a person, interpreting the voice, its inflections, and the vocabulary it uses in social terms. It is, therefore, very important to present callers with a persona that is carefully constructed to represent the brand, not a haphazard set of inconsistent characteristics

• the persona can reinforce and develop the brand in a unique way. Using an automated service

is always a direct experience for the caller and the persona is the only aspect of a brand that customers interact with, and may even develop a relationship with. This makes it the ideal medium for transmitting positive marketing messages

• research has shown that callers will respond more positively if they sense that the “person”

that they are talking to has the right personality for the job. If the persona is developed in the same way as the rest of the brand, using a profile if the typical customer, callers will have a higher opinion of the automated system and will be more inclined to co-operate with it. They will interpret any problems they have in using the system in the light of its personality and be more lenient. They will generally have a much more positive experience of the call.

Devising a persona

Devising a persona for an automated system typically involves finding a human equivalent of the company brand. The designers work to create a character with a name, a clearly defined function, and a mini-biography. This information is used

• by the voice talent agency to select a voice artist who matches the specifications

• by the voice artist to help him or her maintain a consistent character

• the dialog designer to ensure that the application uses appropriate language and performs

tasks in an appropriate way.

Page 91

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 91 of 116

Using persona with SAB

To implement a persona:

1. Record the prompts using the appropriate voice artist and prompt wording.

2. Create the prompts by specifying an existing file (see page 84).

3. Assign the appropriate persona type (see page 86). SAB then stores the prompt in the SAB database with the appropriate persona label.

Page 92

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 92 of 116

The Component Workbench

The Component Workbench is used to import new SAB components. The components are supplied as a .jar file.

Importing components

To import new components:

1. Expand the Component Workbench (by clicking on the + sign) in the Browser window.

2. Click on Import New Components. SAB displays the Import/Update Jar File dialog window.

Figure 89. Import/Update Jar File dialog window

3. Click on the No File Selected button to navigate to and select the .jar file you want to import. The name and extension of the .jar file are displayed in the Name and Document Extension fields; a description of the components is displayed in the Description field.

4. Select any dependent files.

5. Click the Import Document button to read the .jar file and import the components. The dialog box lists the components and the updates that are being imported, and any error or warning messages.

Page 93

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 93 of 116

The Administration resource

The Administration resource is used to manage:

• Dialog Configuration, including language and persona

• Multiple users

• Permissions.

Dialog Configuration

Generating a .war file

For more information about war files and the deployment of SAB dialogs, see Configuring the Speech Applications Builder for Deployment.

A .war file is a zipped archive of files that contains an application and all its resources. It contains everything needed by the runtime part of the SAB platform. Once the .war file has been generated, deployment is simply a matter of installing it on the servlet engine of choice. To generate a .war file:

1. Expand the Administration tree.

2. Click on Dialog Configuration. SAB displays a Generate War File for Deployment button in the Dialog Modeler window.

3. Click on the Generate War File for Deployment button. SAB displays the Dialog deployment war file wizard.

Figure 90. Dialog deployment war wizard 1

4. All SAB applications have a top level flow - the flow that will be run first when a caller interacts with an application. Select the top level flow.

5. Click Next.

Page 94

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 94 of 116

Figure 91. Dialog deployment war wizard 2

6. Either check the version number you want to deploy or check AutoUpdate.

7. Click Next.

Figure 92. Dialog deployment war wizard 3

8. Check the environment you want the war file to run in:

• Tool - under construction

• Production - for deployment.

If you have created a custom environment - such as Testing - it will be presented as an option here.

Click Next.

Page 95

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 95 of 116

Figure 93. Dialog deployment war wizard 4

9. Configure the Servlet for the system on which the .war file is to be deployed, as follows:

Parameter

Description

Adaptor style

Select the type of voice platform the application will be running on.

Audio channel

Select the correct telephony.

Persona

Select the persona to be used - together with Language, this specifies the prompt set to be used

Language

Select the language to be used - together with Persona, this specifies the prompt set to be used.

Grammar set

Select the correct grammar set for the voice platform you are using.

Dialog refresh time (ms)

Specifies how often the system must check the database for updated flows (only if the .war file is generated with Auto-update selected).

Defaults to 600,000 ms

Servlet name

Defines how the SAB servlet will be named to the application server.

Accept default.

Servlet URI mapping

Defines what URI the application server should map the servlet to.

Application servers typically append this to the name of the context SAB is executing in, which in turn they usually name after the WAR file from which it was loaded, without the .war suffix. This means that the default value of /servlet/MainDialogFlow will typically make SAB available at

http://Machine:Port/sab/servlet/MainDialog

if the SAB WAR is called sab.war and Machine and Port are the address and port on which the servlet container is listening for HTTP requests.

Accept default

Page 96

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 96 of 116

Parameter

Description

Application Server

Select correct application server for the voice platform you are using

10. Click Next.

Figure 94. Dialog deployment war wizard 5

11. Configure the servlet to use the SAB database.

These parameters are initialized to whatever values are currently in the ccp.properties file, which correspond to the values the SAB Configurator is using to access the repository database. The values will typically not need to change if the runtime is to use the same database as the Configurator, however it may be necessary to change the address of the database server to a format which resolves correctly from the machine where the runtime is to be deployed.

For example, using the address localhost will only work if the database server is on the same machine as the runtime.

Parameter

Description

Is Datasource

Determines whether SAB should interpret the database URL specified later as a JDBC connection string or as a Java datasource name. Using a JDBC connection string results in SAB making a direct connection to the specified database with that connection string. Using a datasource name results in SAB delegating the connection to a Java datasource defined within the application server.

Database name

Used by SAB to look up a properties file containing low level information about how to use the database.

Database Url

This is interpreted either as a JDBC connection string or as a Java datasource name depending on the setting of the datasource flag.

Page 97

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 97 of 116

Parameter

Description

User

This is the SQL username used to authenticate the connection to the repository database. The required value is entirely dependent on how the database itself is configured.

The user is set by default to sysdba.

Password

This is the SQL password used to authenticate the connection to the repository database. The required value is entirely dependent on how the database itself is configured.

The password is set by default to masterkey.

Cache Directory

When SAB reads flows, prompts and components from the repository database, it caches them in a directory or folder in the filesystem. This setting specifies the location of that cache directory.

12. Click Finish. SAB displays the Save dialog box.

13. Browse to the appropriate location and click Select. Generating the .war file may take a few moments.

Language

When creating a prompt (see page 86) you are asked to specify its language and SAB labels the prompt file accordingly. This feature allows you to create a single application in multiple languages.

Some languages have multiple variants or 'locales' (for example, US English, UK English) which are pronounced differently. When you specify a language with multiple locales, ensure that you use the complete language id, including the correct locale, for example, 'en-US' (for US English), or 'en-GB' (for GB English). The locale setting must match one of those that is supported by the recogniser you are using. The precise id used to specify the locale is recogniser-specific.

SAB’s list of languages and locales is configured in the Administration resource.

To add a language:

1. Expand the Administration tree.

2. Click Language. SAB displays the Language tool in the Dialog Modeler window.

3. Click on the Add button. SAB displays the Input dialog box.

Figure 95. Input dialog box

4. Enter the language name.

5. Click OK.

Page 98

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 98 of 116

The new language is listed in the Language tool and will appear as an option in the appropriate Prompt creation dialog box.

To remove a language:

1. Open the Language tool (see above).

2. Select the language name.

3. Click on the Remove button. The language is deleted.

To edit a language:

1. Open the Language tool (see above).

2. Select the language name.

3. Click on the Edit button. SAB displays the Input dialog box (with an empty field).

Figure 96. Input dialog box (used to edit Language)

4. Enter the corrected language name.

5. Click OK.

Persona

When creating a prompt (see page 86) you are asked to specify a persona and SAB labels the prompt file accordingly. This feature allows you to create a single application that uses multiple personae.

Persona is discussed in detail earlier.

SAB’s list of personae is configured in the Administration resource.

To add a persona:

1. Expand the Administration tree.

2. Click Persona. SAB displays the Persona tool in the Dialog Modeler window.

3. Click on the Add button. SAB displays the Input dialog box.

Figure 97. Input dialog box

Page 99

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 99 of 116

4. Enter the persona name.

5. Click OK.

The new persona is listed in the Persona tool and will appear as an option in the appropriate Prompt creation dialog box.

To remove a persona:

1. Open the Persona tool (see above).

2. Select the persona name.

3. Click on the Remove button. The persona is deleted.

To edit a persona:

1. Open the Persona tool (see above).

2. Select the persona name.

3. Click on the Edit button. SAB displays the Input dialog box (with an empty field).

Figure 98. Input dialog box (used to edit Persona)

4. Enter the corrected persona name.

5. Click OK.

Page 100

Speech Applications Builder Configurator User Guide • May 15, 2004 • page 100 of 116

Debugging a dialog

The Speech Applications Builder provides facilities for analyzing and debugging a dialog flow.

Dialog Analysis

The Dialog Analysis facility lists and describes every path through the selected dialog flow. The data it provides can be used to build test cases.

To view the Dialog Analysis window:

1. Open the dialog flow.

2. Check out the dialog flow.

3. Select the Dialog Analysis tab. SAB splits the right-hand window into two areas, then displays the entire dialog flow in the upper window and lists the individual dialog paths in the lower window.

Figure 99. The Dialog Analysis window with Process Diagram selected