Speech Recognition Condition

Overview of the Speech Recognition Condition

For an Overview of Conditions, click here.


The Speech Recognition Condition can be part of any frame created with Story Flow. It has the same behaviors and properties as any other condition. Unlike other conditions, it contains a recognizer.  

It can only have one recognizer. This makes it easy to use. Consequently, a more complex verbal interaction would require many frames with many Speech Recognition Conditions to work. 

An example of where this condition makes sense is when a verbal acknowledgment or statement is part of a larger procedure. This might occur in a pre-flight checklist before takeoff or confirmation during a delivery.


Required Fields:

Field Name Description
Recognizer A Recognizer can either be set to recognize a keyphrase or an intent.  For a full description, click here.


A Recognizer can either be set to recognize a keyphrase or an intent.

A Keyphrase is a particular word or short phrase to which the script will listen. This keyphrase can be typed directly into the script by a Storyflow author. This is useful for simple voice interactions such as a greeting or an acknowledgment.

An Intent is used for more verbose language. It can use variations on wording to understand what the user is intending to say.

The Intent inside of a Recognizer contains two fields:

  • Intent - This is an Intent catalog item that has been previously defined. The item will already be associated with a voice assistant, like IBM Watson, and a list of phrases/sentences which relate to the meaning for which the system is listening. As an example, the Answer Questions intent is listening for the user to ask "Do you have any questions?". It will also accept variations on this like "Have you any questions at this point?" or similar. Intents allow the user to speak naturally. They let the system infer what the user is aiming to say.
  • Confidence - The level of confidence that the user has said something related to the Intent. This is a value between 0 and 1. It can be thought of as a percentage. So, 0.6 = 60% confident that the user has said something related to the intent. The higher the level of confidence, the more precise the user's language must be for the Intent to be recognized.