Business Customer Support
Ongoing Business Customer Support Services
Conceptual and Pre-Development Support
Deciding when and how to implement speech recognition in your product is a big decision. Additionally, the human interaction factor is a critical element of speech recognition interface design.
How a speech solution is designed, from the human interaction stage through the accuracy stage, and how the results are processed into meaningful results to your customer, require critical planning and human factor design support, in addition to, expertise in the actual speech recognition design.
All of these factors play a critical role in the end result perception your customer has on the speech interface that is designed and deployed.
MatrixHCI has years of experience in the pre-design and architectural stages of speech interface design, and can help you formulate the exact product design for your customers as you are still in the exploratory stage of your process.
Development Support
Once your system is conceptually designed, our professional and experienced team will be there with you as we design and develop your custom speech application.
Our collective team development capabilities also bring tested and proven Software Engineering principles that ensure solid architectural software design, human factors engineering, and software design implementation processes that ensure the development of a successful software solution.
Our agile approach to software design also ensures that the development process is developed in iterative cycles where you can review the progress and provide input as the project is developed.
Our iterative development process allows for changes during development should needs in the market change in a way that necessitates change in your customer applications.
Post-Development Support
MatrixHCI is also with you during and after the release of your product in order to help make any changes, modifications, and possible updates to your solution as the market progresses and matures.
Additionally, speech recognition training and accuracy is an ongoing process where audio from your actual system is routinely collected and used for future training and refining of your acoustic models and language models in an effort to continually improve recognition accuracy.
Initial acoustic models for first-time products are usually trained with what is called “out of domain” audio which is audio collected in studios or used from preexisting recorded and transcribed audio corpora data.
Once your product is functional and collecting audio from within the confines of your end-to-end system, this audio is used to further train the baseline acoustic models that are initially trained and utilized in the initial product.
Audio collected from within your complete system is called “in domain” audio, and its value is more valuable than “out of domain” audio. It is estimated that “in domain” audio is by a factor of 10 times more effective in training and improving your initial acoustic models.
MatrixHCI is there with you for the continued collection and retraining of your acoustic models as this “in domain” audio is collected from your customer use and becomes available.