Chapter 1. Introduction to Voice User Interfaces

The following is a true account of the first experience of one of the authors (MC) deploying a commercial spoken language system:

The first application we ever deployed at Nuance was a voice-driven stock quote system for Charles Schwab & Company. The months leading up to the deployment were very exciting for me personally. I had spent more than ten years in the research labs at SRI developing spoken language technology. Finally, we had started a company to commercialize our technology. We had signed our first customer and were almost ready to deploy our first system. We were about to prove to the world that this technology had real value that it was ready for "prime time."

I observed the early usability tests. I was behind the one-way mirror as subjects came into the test room. I watched them pick up the phone and use the system to get stock quotes and set up watch lists and watched as the experimenter questioned them about their experience.

The first subject arrived. He was an 83-year-old man from San Francisco. He came in the door, hesitated, and said, "Oh no, I forgot my hearing aid!" Five seconds into the first real user test, and we were already hitting problems we had never anticipated in ten years of research!

The test proceeded. The subject had a great deal of trouble using the system. He had difficulty hearing the prompts, was confused about what he could say, and got lost in the application. By the end of the experiment, he had never succeeded in setting up a watch list and never received a quote on a company he requested.

At that point, the experimenter asked the subject about his reactions to the application and whether he thought he would use such a system. He loved it! He described the experience as the first time in years he had been able to have a conversation with anyone so patient, so willing to repeat themselves, and so willing to talk with him and never get frustrated. He promised he would be calling this system every day!

This story illustrates the two key themes of this book:

1. Understanding basic human capabilities is key to the design of effective user interfaces.

2. Understanding the user's needs and goals, in the context of the business goals, is key to the design of successful applications.

The two themes differ in that the first addresses general human capabilities: understanding what is easy and what is hard for all (or most) people and how to exploit this knowledge to optimize design choices. (Here we are concerned with human cognitive capabilities and linguistic behavior rather than individual physical capabilities such as hearing, as suggested by the story.)

In contrast, the second theme specifically addresses the application at hand: what you need to understand about the intended users, the task, and the business goals in order to focus the design on meeting user and business needs simultaneously. In the story, we could have considered the system test a success if merely providing an electronic conversational partner were an acceptable result. The test result did not, however, meet the business goal of providing value with respect to the handling of brokerage accounts.

In The Humane Interface (2000), Jef Raskin draws a distinction between user-centered design and human-centered design. He describes user-centered design as a design process focused on studying the task-related needs of the intended users of a specific application. He then goes on to emphasize the importance of human-centered design, which is focused on "making sure that the interface design accords with universal psychological facts." He argues that it is more important to understand the cognitive capabilities and limitations of humans in general and to apply that understanding to interface design problems.

In this book, we argue that both human-centered and user-centered design are key to the creation of successful user interfaces. We spend substantial time explaining the core principles of design based on an understanding of human cognitive capabilities and human linguistic behavior. However, we also examine in detail the methodologies by which you can integrate an understanding of the intended users, the application tasks, and the business goals of the system to create effective interfaces that provide successful and satisfying user experiences.

Our focus is entirely on voice user interfaces. Moreover, given the ubiquity of the telephone and the large number of spoken language systems currently being deployed for over-the-telephone use, we focus specifically on voice user interfaces designed for the phone.