Creating a Speech Application Specification

Use an application specification to develop and refine the features of an application. A speech application specification should contain the sections described in this topic. A template is provided in Speech Application Specification Template to make it easier to quickly create a specification.

For a description of a deliberate and disciplined approach to managing technology projects, see Microsoft Solutions Framework (MSF) in the MSDN Library.

Target Organization

Begin a speech application specification by describing the organizations that might use the application. Typically there is just one target organization. This section of the specification should focus on a detailed analysis of the business that the designer and developer are building the application for. The following questions should be considered.

• Provide an overview of the business. For example, what the business do and how does it do it?
• Why speech? Why is the business looking into speech? What do they hope to get out of the application? What are the success metrics?
• What persona does the organization project? What values or attributes does the company want to project?
• What's the schedule?

For an example, see Specifying the Target Organization

Target IT Persona

Describe the specific characteristics of individuals that might drive an organization's purchase of this application. This section is focused around understanding the needs and desires of the person who will be administering, monitoring, or managing the application.

For an example, see Specifying the Target Information Technology (IT) Persona

Target Users

The application specification should provide in-depth research describing the intended user of this application.

• Basics: Who is the customer? Can they be segmented in some sensible way? What are the priorities?
• Specifics: How do they speak? How do they think? How do they view themselves? What will customers be thinking when they encounter a speech recognition application?

For an example, see Specifying the Target User.

Application Requirements

When you have a solid understanding of the customer and the business, derive a set of requirements for the application. A requirement is some capability the product needs to have. Note the following two sample requirements:

• Must be able to handle twelve simultaneous calls.
• Must be able to resolve problems in the dialogue automatically.

Application Features

Describe and then rank the application's features. A feature is an implementation of a requirement, and can have development and testing costs. The following bullets show the features corresponding to the requirements in the previous section.

• Super Connect Topology
• VUI Doctor Tuning Tool

User Scenarios

Describe the ways the application will be used. Include both end-user scenarios, and customer scenarios and administrative scenarios, and prioritize these scenarios. For an example, see Specifying User Scenarios.

System Persona

Speech applications use a human voice to interact with the user, which gives the application human-like characteristics. To improve effectiveness, base the application's voice on a specific personality. To increase the realism and appropriateness of the personality carefully design, as explained and described in Specifying a Speech System Persona.

Creating a Mental Image

Evidence indicates that we infer personality and social characteristics from the voice of a person, even if we know the voice to be recorded. In effect, we create a mental image of a personality based on the voice and language choices. In speech applications, this mental image is known as a persona.

Businesses are often concerned with their brand image and want that image to project appropriately to customers. An auto-attendant is often the first point of contact for an external caller. Thus, a persona that reflects the brand image is an important consideration. A good speech interface engages the user in a seamless and consistent conversation. A persona can act as a style reference to ensure that the choice of language and the prosody of speech remain consistent throughout the call.

For an example, see Specifying a Speech System Persona.

See Also

Speech Application Design Considerations | The Speech Project Lifecycle