Using the Voice Input Resource

Owned by Sara Johnston
Last updated: Jul 15, 2022 by Catherine Hillier (Unlicensed)
2 min read

For an overview of Conversational AI, click here.

The Voice Input resource allows you to author a script that listens for more than one intent at a time and author an outcome when none of the intents or keyphrases are recognized.

To use the Voice Input Resource:

Authoring an Intent Recognition Success

  • Drag a Voice Input resource into the resources area of the frame

  • The prompt will add guidance/a persistent message to the screen to help the user

  • Click the '+' beside options to add a keyphrase or intent

  • Select the event that will fire when the intent is interpreted successfully (can be any of the predetermined events in the list, or you can add a custom event)

  • Click the '+' beside the event to create a sub-frame or branch that will open when the event fires

  • Add the resources/actions you want to happen when the intent is recognized to the child frame

Watch this video to see how to author a greeting to a character in an office scene:

Authoring an Intent Recognition Failure

You can also detect when none of the intents in the resource have been recognized using the “Fail” intent. This helps give the user feedback that they are not being understood. It can also serve as a trigger for offering up more guidance to help the user stay or get back on track.

  • Click on the blue Event Link Shortcut to bring up the list of events

  • Select “Fail”

  • In the new branch that is created, add the resources/actions that you would like to occur on failure of the Voice Input resource

Note: In this video, the author did not toggle the “ignore previous” flag on the fail condition - which leads to an infinite loop. make sure you check “ignore previous” on the fail condition!

In the Headset