TeleFlow Step 0050
Collects a selection through a touch tone or speech recognition, following the path related to the user's selection.
Speech recognition is only available if an SR Acquire step was previously executed. Without SR, this step will accept Touch-Tone input only.
When this step is placed on the flowchart, most of its related Action steps will also be created. Any Action steps that are not required can be hidden through an option on the right mouse button menu, or by pressing Ctrl-H on each. If the user should enter an option that is hidden, the Server will follow the path of the All Others Action step. The four Action steps for the letters, and , are automatically hidden. To use any of these four action steps, you must unhide them from the step's right mouse button menu.
Variable to receive selection: The touch tone digit or speech recognition result will be sent to this TeleFlow variable. This field is blank by default, but may be any TeleFlow variable. Note that no result variable needs to be collected by this step; collecting a result variable with this step is optional.
TIMEOUT (ms): This field is the maximum time, in milliseconds, before this step will time out and follow Time Out path if no touch tone or speech entry is detected. If this field is blank (the default), @TIMEOUT will be used.
Speech Recognition Options: These options are only necessary/used when speech recognition is enabled. If they are all left blank in a speech recognition enabled system, the step reverts to using touch-tone only behavior.
THRESHOLD confidence %: The result confidence is compared against this minimum confidence to determine a valid result. This field is populated with 50 by default. This field must not be blank.
Variable to receive confidence %: The result confidence will be put in this TeleFlow variable. This field is populated with @SR_CONFIDENCE by default. This field can be blank.
Remember this recognition for verification through SpeechTrainer: If the TeleFlow SpeechTrainer is active, checking this box ensures the recognition will be available for review in SpeechTrainer.
Variable to receive SR Handle: Use this to capture the SR Handle for the utterance if the application uses the SR Flag Result step to flag certain types of SR results in SpeechTrainer. Utterances flagged in SpeechTrainer can be found more easily, and used to isolate specific problems the application is set up to identify.
Retain audio on Threshold: Check this checkbox to compare the speech recognition audio against another grammar (such as SR Get Yes/No or SR Get String) when the result confidence is less than the THRESHOLD confidence.
Base grammar (name or file): This field specifies the name of the Nuance pre-compiled grammar, or the SRGS grammar file for use with LumenVox. If this field is left blank, only the additional grammar text will be used for speech recognition. If both this field and the additional grammar field are left blank, this step will respond to touch tones only.
Additional grammar text: This field specifies SRGS grammar rules in addition to the base grammar, but can only be used by LumenVox. If this field is left blank, only the base grammar will be used for speech recognition. If both this field and the additional grammar field are left blank, this step will respond to touch tones only.
through , , through : A touch tone was detected, and the confidence was set to 100, or the speech recognition grammar returned a result matching the number indicated with adequate confidence. (E.g. The path would be followed if a caller presses "0", or if the result from Speech Recognition was "0" and the confidence was equal to or higher than the THRESHOLD confidence %)
All Others: A touch tone was detected that did not correspond to a connected action step, or a speech recognition result was detected with adequate confidence that did not correspond to a connected step. Use the variable to receive selection to determine the result.
Unrecognized Voice Action: The confidence was less than the THRESHOLD confidence.
Time Out: No speech or touch tones were detected before the No-Voice TIMEOUT elapsed.
Special Hardware Considerations
- Dialogic: Only touch-tone and prompted SR are supported (I.e. SR with barge-in is not currently supported).
- AudioCodes (Ai-Logix): Touch-tone support only.