• Usability Surveys To Test Speech Applications
• Avoiding The Pitfalls of Speech Application Rollouts Through Testing and Product Management
• Developing Speech Dialogs: A Designer's Perspective
• How "Natural" Should a Voice Interface Be?
• Application Design: Error Handling
• Building Applications with Alternative User Interfaces

Avoiding The Pitfalls of Speech Application Rollouts Through Testing and Product Management (TD-04)
Speech self-service applications can truly become the voice of a company. Unfortunately, it only takes one bad experience for a customer to stop using self-service applications. As a result, application performance can have a significant effect on customer loyalty and a company's image and bottom line.

Developing Speech Dialogs: A Designers Perspective (TD-07)
Changing technology allows for the development of innovative ways to build new man-machine interfaces. In particular, adding speech to the mix opens up a host of new possibilities. However, the fast-changing technologies also create a lot of confusion with respect to where and how speech can be applied and what technologies and tools are needed to build effective applications. This presentation will attempt to answer some of these questions from an application development perspective.

How "Natural" Should a Voice Interface Be? (CCA-02)
Since the inception of natural language processing (NLP) technology, its proponents have promised a revolution in voice interfaces for call centers, predicting a future in which callers will interact with IVR systems as if the systems were human agents. Exciting stuff, but is this really what you or your callers want? There's a reason why customers sometimes seek that human touch: because there's a real live person on the other end. Regardless of how lifelike an NLP-enabled IVR interface may be, it's still a machine and will never be as equipped to handle the more complex issues as a live agent. Plus, callers may become angry if they feel they've been tricked and the person they've been talking to is, in fact, a machine. But when a caller does have a simple request or transaction to complete, the less-than-human interface typically offered by today's IVR systems may be just the ticket. There are advantages to interacting with a machine: hold times are generally shorter; the caller doesn't have to make small talk with a live agent; and simple transactions can be handled quickly and easily. NLP does, in fact, have a definite place in the call center — companies simply need to learn how to leverage the advantages and sidestep the pitfalls.

Application Design: Error Handling (TD-10)
While application designers and recognition experts rightly measure the success of applications by such parameters as “did the user accomplish his or her task” and “how accurate is the recognition,” designers also need to consider how much the user enjoys using the system. However, we have found this does not correlate directly with how accurate the recognition is, but rather, how easy it is to use the system and how elegantly the system handles errors. As much as we wish it were, speech recognition is not perfect. The user often will not perceive whether an error is caused by their selves, the environment, or recognition engine. Because of this, the system will not always behave the way the user thinks it should. Keeping in mind branding and persona issues, this presentation will offer tips and real-life, case-study examples on handling different types of user and application errors that if utilized, create a more robust, helpful voice application that allow users to feel as if the system is working with them instead of against them, and will ultimately leave the user with a good impression of your company’s customer service. Dialog and programming tips for error handling include Navigation, No Matches, No Inputs, Multiple Fields of Information filled with one utterance and will address the ever-debatable topic “Do we apologize for making errors?”

Building Applications With Alternative User Interfaces (MD-02)
Soon, users will want to access business applications using new mobile devices with multimodal capabilities.  Come learn how to support multiple alternative user interfaces, including Web-based GUIs for PCs, voice-only for telephones and multimodal for the new mobile devices.  Discuss when it's possible to author a single user interface that can be used for all three types of devices, and when alternative user interfaces must be used for each device.  Discover architectures in which a single application can support a visual user interface for PC users, a verbal user interface for telephone users and a multimodal user interface for the users of mobile devices.  Evaluate languages for developing verbal and multimodal applications within this architecture, including SALT, X+V and the emerging W3C multimodal languages including EMMA and InkML, as well as the Systems/Environment specification.  You'll understand the various approaches for providing user interfaces, and be able to recommend a corporate strategy for supporting user interfaces for multiple devices.