Flylib.com

Books Software

 
 
 

6.5 Conclusion


6.5 Conclusion

VUI users know that the prompts they are hearing are only recordings. Some users even know something about speech recognition software, IVR platforms and networks, and backend integration. But even these tech-savvy users are "programmed" to infer mental images from the voices and linguistic choices they encounter. Designers therefore need to be "in control of the image that their application projects so that users will favorably assess the company's brand, values, and support of them as customers." [2] Whether or not you have planned for it, your application's persona is an expression or statement of an image. Don't leave users' perceptions of your application, and therefore of the company, to chance.

[2] Margaret Urban, personal communication.


Chapter 7. Sample Application: Requirements and High-Level Design

Albert Einstein made the comment, "Example isn't another way to teach, it is the only way to teach." Indeed, when learning VUI design, you can read lists of procedural steps and guidelines, but, until you see the process in action, you can never fully comprehend the details, subtleties, and constraints imposed by real-world problems.

This chapter begins our presentation of a sample application. Throughout the book, we will carry on this example, stepping through all the details of definition, design, development, and deployment. You will have a chance to see the methodology and design principles in action. Keep in mind that our purpose is not necessarily to advocate a particular solution or set of design choices for the application at hand, but rather to show the process in action. We have fashioned our design decisions and test results for this example to demonstrate particular points. They do not necessarily represent actual test results on a real deployment.

The sample application is for our imaginary client Lexington Brokerage. The application it wants to build will allow callers to get stock information, place trades, and manage their accounts. We begin with a brief introduction to Lexington Brokerage and its goals. Then we dive into the details of requirements definition and high-level design. [1]

[1] The material in these chapters on the Lexington Brokerage application is based partially on presentations by Tony Sheeder and by Rebecca Nowlin at Nuance V-World: "VUI Design Under the Microscope," V-World 2001.


7.1 Lexington Brokerage

Lexington Brokerage is a small firm, not terribly well known, but aggressive and eager to become a major player in its industry. In a competitive analysis, Lexington executives have discovered that other brokerages have been offering speech recognition, instead of standard touchtone systems, in their call centers. They have met with industry analysts, who suggested that not only have the systems provided significant cost savings, but also brokerage customers are very satisfied with them. The analysts reported that customers prefer speech to touchtone systems and that they appreciate the easy access to their accounts and up-to-the-minute information on the performance of their portfolios. One analyst described a survey he performed in which customers of brokerages with speech offerings rated their brokerages higher on customer service.

Lexington has a touchtone system that provides stock quotes and basic account information. The executives have decided to create a speech system in order to expand the firm's automated offering to include trading and portfolio information, and to provide customers with an easier-to-use, more intuitive interface. They expect the automation of telephone trading to lead to significant cost savings.

We have just signed a contract to design, develop, and deploy the system. Lexington wants to deploy quickly, although it is concerned with achieving a high-quality result and making the deployment process smooth and predictable. We enter the first phase of our project: requirements definition.