Speech Tuning & Voice Apps

Speech-enabled messaging facilitates efficiency, and Adomo seeks to optimize that efficiency by improving the system’s recognition software’s ability to identify and then execute what speakers request. In order to derive the highest accuracy from the speech recognition software, it is usually necessary to adjust certain aspects of the system, e.g. recognition dictionaries and/or recognition engine parameters. These adjustments, or application "tuning," require a specialized procedure that progressively improves the application's recognition of callers' speech, and therefore enhances the user experience.

As a primary input mode for Adomo Unified Messaging, successful recognition of speech contributes to a satisfying user experience. However, the true indication of a successful interaction is if a user achieved what she or he intended. Callers to the auto attendant need to get to the person they want to reach (or leave a message), and voice messaging subscribers need to successfully voice address messages and send them with delivery options, if desired. During the tuning process, Adomo will monitor and evaluate interactions such as these that occur between the system and speakers. Based on the results, we will determine if tuning or configuration adjustments are needed to improve recognition accuracy, and thus maximize the occurrences of task completion, or "completion rate."

During the Speech Tuning engagement, Adomo professional services engineers increase the already high level of name and word recognition by:

•  Determining your tuning needs
  
  
•  Examining logged speech data, evaluating names and commands that are not recognized
  
  
•  Collecting subjective user feedback using online forms