The voice solutions market is fueled by the rapidly growing number of consumers who want easy anytime, anywhere access to information and services via any device - wireline or wireless. With more than 1.5 billion phones and over 800 million mobile device users and 1 billion landline accounts worldwide, the telephone has become the ultimate access device to the Internet.

Enterprises are looking for new ways to reach the consumer as well as extend the information to their employees. This combination is creating a powerful force that is driving the voice business market. Additionally, vast improvements in advanced speech recognition technologies have brought new applications into mainstream use as enterprises seek to increase revenues while reducing costs. In fact, the Kelsey Group has estimated that the U.S. market for voice-activated applications and infrastructure targeting both the enterprise and carrier markets will hit $2.2B by 2006, with open standards-based solutions overtaking proprietary voice response platforms by the end of 2003.

Our Development Challenge
As enterprises face increasing cost pressures and the need to generate additional revenues through product and service differentiation, new solutions are needed to deliver consistently high-quality customer service while improving operational efficiencies. Customer service through recent investments in Web technology tackle part of the business challenge, but this approach doesn't help customers who do not have Internet access or for whom Internet access is inconvenient or impractical.

Interactive Voice Response (IVR) products have been available for many years to automate call processing to alleviate some of the efficiency pressures affecting businesses. But until recently, these products were based on touch-tone or Dual Tone Multi-Frequency (DTMF) technology. Users often find touch-tone technology cold and unfriendly, and in general it leads to long and confusing menu structures that are not at all intuitive. Moreover, these products typically operate in a proprietary environment that requires the developer to tie learning to a specific voice response platform that doesn't leverage an understanding of Web technology. Now there are solutions that deliver the benefits of speech recognition and text-to-speech within a VoiceXML environment and can increase flexibility and simplify deployment and integration tasks.

Design, Develop, and Deploy Speech Applications
As developers we need easy-to-use and open standards-based tools that allow us to quickly and easily build advanced VoiceXML-based voice user interfaces through the dynamic generation of reusable code. These development tools do exist, and they provide the foundation for a powerful, market-proven application development environment for the creation of voice portals with speech recognition and text-to-speech applications.

The capabilities of today's VXML-based development environments now incorporate built-in, icon-based tools - from code editors and graphical dialog designers to grammar specifications and rehearsal and emulation tools - that measurably reduce the time and complexity of development, coding, and integration for customers' voice solutions. These capabilities allow you to concentrate on application call flows and the user experience, rather than on cumbersome code syntax and structure. However, when selecting a voice automation tool to begin work, care should be taken to ensure that the tool is fully compliant with the latest VXML specification approved by the WC3 and that you can capitalize on the benefits of packaged application components to speed up deployment.

The Speech-Enabled Application Developer Challenge
One of the biggest challenges I've faced in developing speech-enabled applications is in creating reusable components with enough flexibility in the Voice User Interface (VUI) to be truly reusable. All too often I've rewritten code that performs functionality that has already been written, but has to satisfy different VUI requirements. What I've learned is that though there are consistent reprompting and error-handling strategies, there is no universal VUI. Often each company, and even each department within a company, has very different VUI requirements based on the end users of the application. For example, a Personal Identification Number (PIN) reset application has basically the same functionality, but depending upon whom it is intended for, there are differing security considerations and user experience level considerations that drive what information is required from the end user and how the prompting is structured to verify identity. In reality the information gathering for user verification is a VUI requirement, but the PIN reset functionality is basically the same in all situations. The is the niche that packaged applications can fill by providing fully configurable packages, both in VUI and data access, that can either be used as standalone applications or be incorporated easily into larger applications.

Packaged Application Components Speed Deployment
To deliver more active and effective customer service, while at the same time leveraging your knowledge of XML data, I recommend using new-generation packaged voice application components. Why? Packaged voice application components provide real-world feature functionality that literally transforms pervasive service and product delivery processes into conversations.

While some voice solutions are based on the strategy of creating vertical applications for specific market segments, you can now also leverage “horizontal” packaged application components that address the common functions needed in many industries. As you know, password resets, for example, are much the same in any market application. But these reusable modules, deployed within larger applications or with customized feature sets, will allow you to leverage the efficiencies of packaged components while delivering the unique solutions your market requires. In fact, you can deploy packaged application modules to manage crucial speech-related functions such as: