- How We Help
- Thinking About Speech Technologies
- Stages to Success
- Figure: How to Derive Project Requirements
How We Help
Disaggregate provides consulting for speech technology projects. Consulting is available on an ongoing or as-needed basis to plan projects and help resolve project issues. We also provide dedicated in-house consulting to carry a project from design to deployment. We specialize in telephony applications.
If you choose ongoing or as-needed consulting, you obtain your own in-house "speech guru." We provide the specialized expertise and deep knowledge of speech technologies: How speech technologies interact with the rest of the project, how they affect the system architecture — and how to make the project successful.
If you choose in-house consulting, we provide a system architect for speech technologies to develop project requirements and drive the requirements through the design, implementation, testing, and deployment stages to ensure that the project meets expectations.
Thinking About Speech Technologies
When people talk about "speech technologies," they usually mean the fundamental, underlying technologies:
- Automatic Speech Recognition (ASR)
- text-to-speech (TTS)
- speaker verification and speaker identification, also known as speech biometrics
These technologies provide the basic building blocks, but speech technologies can't simply be added on as an afterthought to a project. Speech technology projects require careful planning and thorough preparation.
Stages to Success
Disaggregate provides an expert, experienced system architect. At each stage, the architect takes definitive steps to guide the project to a successful conclusion:
|
Investigate |
The architect works with managers, marketing departments, end users, and others to determine what the project is suppose to accomplish. He decides whether or not the goals are attainable and proposes speech technology solutions. |
| Educate | The architect educates managers and technologists about the proposed solution, along with opportunities and risks. He outlines the responsibilities of non-technical organizations such as marketing departments or call center divisions. The architect must be available to discuss the project with top-level managers. |
| Instantiate | The architect creates the requirements for the project and passes them along to the system designers and test organizations. The architect attends design reviews to ensure the requirements are properly understood and fully met.
The architect mediates between your organization and its speech technology vendors to make certain that all sides clearly understand each other and have a shared set of expectations. Perhaps no other detail is more important than good "human factors" — designing the user interface to prevent user errors and user frustration. Prompts must be worded correctly, recognition and user errors must be handled correctly and unobtrusively, and the application must be thoroughly tested. |
| Evaluate | The system architect works with the application designers, hardware engineers, test teams, and others to make certain that their designs, prototypes, and tests are properly aligned with requirements. Prototypes and small-scale trials uncover how the system and the users perform in the real world. |
| Update | Speech technologies include monitoring to determine how the system works in practice and to catch hardware and software errors. After deployment, the team uses feedback from performance monitoring to improve results by modifying prompts, grammars, and other variables. |
Figure: How to Derive Project Requirements
