Voice Recognition Troubleshooting

1. TRAINING
Proper training is a key component to optimal voice recognition results. Make sure the customer trains the phone under the following conditions.
(a) Engine idling
(b) Fan speed on low
(c) User in drivers seat in normal driving position with head forward towards microphone.
(d) User speaks in a clear, normal voice in response to the prompts.
A noticeable improvement in the recognition rate can often be made by having the user retrain the phone after using it for several weeks.
By this time he/she is more acclimated to speaking to the phone and has settled into a pattern of speaking the directory name tags.

2. PROBLEMS ENCOUNTERED
(1) Access Denied Message Given - Service level restrictions may be in place; set service level to 4 for full use.
(2) Name Substitution - Substitution occurs when names are too similar for proper distinction. Have the customer change problematic words to make them more distinguishable from each other. Avoid one syllable words.
Names like "Karen's Home" and "Karen's Office" have a higher probability for substitution.
(3) Inability to Access Voice Memory Tags
(a) Poor voice tags-scroll through the voice memory and note those name tag locations which sound particularly weak or unclear; these voice tags could be candidates for retraining for the customer.
(b) Make sure the customer is saying the name exactly how he trained it.
(c) The user must speak louder while keeping the same tone and pitch of voice used during training, especially when background noise is significant (i.e. while travelling on the highway, in presence of loud trucks, etc.)
(d) Make sure the user is in the correct user mode (A or B) when speaking the voice tags.
(4) Customer dislikes the sound of the repeated name - As described in the user's manual, "Thank you" or a beep can be programmed in place of the repeated name.