Device Detection on Angel.com is done through the Outbound API.
An option is provided under the Outbound site properties page to Enable/Disable Outbound Calling and transfer the call to specific voice site pages depending on the results of the device detection. NOTE : You can set the call to be routed to different voice page depending on device detection ( Human or Voicemail )
Note: Barge-in and Site Commands MUST be disabled on ONLY the first voice site (transfered to when a Human is detected) if using Outbound, Device Detection, and Voice Recognition together.
This is important because on an Outbound Call the Voice Site MUST first give instructions and then listen for responses. If the person being called speaks hello followed by something it must be ignored and NOT taken as a response to the Voice Site. Otherwise, you may hear "I didn't get that." or the prompt being repeated.
If you are having problems with the Outbound Device Detection (long pauses after the Outbound Call Connect to the phone number) try the Device Detection Settings on the Outbound properties page.
To understand the how the above settings work please read on -
Case 1 - No Speech Time Out ( No speech input from receiving end )
- Outbound Call is placed ( Ring Ring )
- Receiving end picks up and no one speaks ( No Sound )
- No Speech Time Out kicks in ( If no one speaks within 1400 milliseconds ) - The system assumes that its a Human
Case 2 - Max Listen Time for Voicemail ( Receiving end keeps talking )
- Outbound call is placed ( Ring Ring )
- Receiving end picks up ( Does not Stop Talking )
- System listens for Max 60 seconds and then assumes its a Voice Mail
Case 3 - Silent Time and Record Duration Threshold
- Outbound Call is placed ( Ring Ring )
- Receiving end picks up ( Lets say , Speaks for T1 milliseconds )
- As soon as the receiver stops to speak, and remains silent for the time set in 'Silent Time', it triggers an event
- EVENT
- IF T1 + 1100 ( Silent Time ) > 4500 ( Record Duration Threshold ) THEN receiving end is a VoiceMail OR
- IF T1 + 1100 ( Silent Time ) < 4500 ( Record Duration Threshold ) THEN receiving end is a Human
- Based on the above logic the system would detect if the receiving end is a HUMAN or VOICEMAIL
NOTE - The times specified in the above text is referring to the settings in the Screenshot.
This is a great KB article to explain the outbound settings
contributed by Sameer Padha on Jul 30 7:37am
