ASRAppGenThe Speech Applications Generator


Introduction

The Speech Applications Generator is a web based management interface that has been developed for the creation and maintenance of VoiceXML applications. It facilitates the deployment, by the customer, of business-specific applications, from a library of modular templates. Once an application has been created, it can be configured using the Applications Generator interface. Its purpose is to simplify and abstract the deployment of advanced ASR, (Automated Speech Recognition), applications, by using customisable Template application modules. There is no limit to the number of concurrent applications that can be deployed and configured using the Speech Applications Generator interface.

The use of VoiceXML in the design and delivery of Speech Enabled self-service applications has not only greatly reduced both the time, complexity and associated costs previously linked to their deployment, but also opened it up as a viable and affordable option for smaller organisations.

By incorporating Speech self-service applications into new or existing IVRs, live agent talk-time can be reduced or eliminated for many repetitive tasks. The applications can be deployed either independently, as a stand-alone service, or in conjunction with live agents, whose involvement in a call may range from providing back-up when a caller asks to be transferred, to taking over a call when the self-service steps are complete. Utilising a blend of Speech Recognition and DTMF based IVRs, and interfacing directly or via web service calls to back-end systems, the inclusion of self-service options empowers the caller, offers consistency of caller experience and, when combined intelligently with desktop CRM, reduces 'Average Handling Time' (AHT) in the Contact Centre.

For further information and technical specifications, click on this link: Speech Applications Generator
For an overview of how to Create a VoiceXML Application: Create a VoiceXML Application in 5 easy steps

Call Routing Application Builder

The Graphical Speech IVR Generator is an advanced feature of the Speech Applications Generator. It is an Applications Builder primarily designed to allow the client to create business-specific Call Routing dialogs using an intuitive graphical interface. Individual steps can be added, deleted, and configured, both in the initial design stage and at a later date. For example, if further menu items or options are required, they can be added and once submitted, the associated grammar will be automatically updated. The client has full design control as they can select and configure the steps relevant to their call routing requirements.

In addition to providing call routing functionality, the Generator can link to an existing self-service application, for instance, Payment Capture. Once the self-service application steps have been completed, the caller is returned to the automated call routing flow.

To see how to build your own Call Routing application, watch the screencast: Call Routing Screencast
To view a sample Call Routing call flow: Call Routing Dialog

Voice Biometrics

The Applications Generator supports the use of Voice Biometrics for caller authentication. Enrollment of callers' voiceprints, and subsequent verification against those voiceprints, can be enabled at any point in a Call Routing Application (See screenprint). To create a complete verification application there must be at least two sets of applications - one for training and the other for verification.

biometrics

Training

Voiceprints are stored in a verification database, along with a unique key that is used to select the voiceprint during verification. Voiceprints that go into the verification database are created using a training session. A training session is the process of prompting the user for required phrases to be used for verification and storing the user's utterances in the recognition database. These utterances will be used to create a voiceprint for the user, linked to a unique key.

Verification

A verification session is where a user's voice is authenticated against the voiceprint that is stored for them in the recognition database. For a verification to succeed the user must already have an associated key in the recognition database. This key is created during the training session.

Static Applications

The basic Speech Applications Generator package allows the customer to create business-specific Applications from a library of reusable VoiceXML components and templates, which include Identification & Verification, Payment Capture and Account Status & Management . New applications are fully customisable and can be maintained and managed via the interface; the configuration can be adjusted to fulfil customer requirements. In addition, the package also facilitates the customising of prompts and grammars with an upload data utility allowing real-time changes to scripts and call flows.

For more information about creating or editing a Static application click on this link:
Static Applications


Macaw Speech and using Interactive Intelligence's Customer Interaction Center(CIC)

macawinin