Studio uses Widgets to represent various parts of Twilio's API that can then be stitched together in your Studio Flow to build out robust applications that require no coding on your part.
The Connect Virtual Agent Widget allows you to connect a Twilio Voice call to a Google Dialogflow CX agent.
You can leverage a native integration between Twilio Voice and and Dialogflow CX to provide callers with conversational AI experiences. The integration also allows for operations such as live agent handoff and customer barge-in.
Twilio provides the telephony aspect of the conversational IVR experience and Dialogflow CX enables agent configuration, including setting intents, sentiment analysis, speech models, and agent responses.
To use the Connect Virtual Agent Widget, you must connect your Dialogflow CX agent to Twilio in the Google Dialogflow CX cloud console. Review the prerequisites and steps for the integration in the Dialogflow CX Onboarding Guide.
After you complete the integration, Twilio will automatically create a new Studio Flow containing a Connect Virtual Agent Widget that is connected to your Dialogflow CX agent.
The Connect Virtual Agent Widget requires the connector name for the Google Dialogflow CX instance. You can find the Connector Name in the Dialogflow CX Connector in the Twilio Console after you have completed the One-click telephony integration in the step above.
|Connector Name||The unique name configured in your Dialogflow CX connector instance.||empty|
Editing the Connector Name within the Connect Virtual Agent Studio Widget alone will break the integration between Dialogflow and your Twilio project. If you need to update the Connector Name, you must edit it in the Dialogflow CX Connector and the Studio Widget.
The Connect Virtual Agent Widget also accepts a number of configuration options that you can use to configure agent interactions and analysis.
|Status Callback URL||URL to send status callback events from Twilio||empty|
|Status Callback Method||The HTTP method to use when requesting the statusCallback URL. Accepted values are GET or POST.||POST|
|Parameters||Key-value pairs used to send custom session parameters to your virtual agent to drive personalization and/or parameterize your agent’s response||empty|
|Configurations||Key-value pairs used to override underlying connector properties (e.g. language, sentimentAnalysis, voiceName, welcomeIntent, etc.) and/or modify your virtual agent’s default behavior||empty|
These events trigger transitions from this Widget to another Widget in your Flow. For more information on working with Studio transitions, see this guide.
|Completed||The Dialogflow CX agent ended the connection because an intent matched the "end of the conversation" intent indicating the call completed successfully|
|Live Agent Handoff||The Dialogflow CX agent returned a live agent handoff response indicating the call needs to be escalated to a human agent|
|Hangup||The caller hung up during the Dialogflow CX agent interaction|
|Failed||An error occurred during VirtualAgent processing|
When the Connect Virtual Agent Widget executes, it will have stored the following variables for use throughout your Studio Flow (where
MY_WIDGET_NAME is the name of your actual Widget). For more information on working with variables in Studio, see this guide.
|Name||Liquid Template Language|
VirtualAgentErrorCode will be present if status is failed (if the Widget ends through the Failed transition).
VirtualAgentProviderError may also be provided if
VirtualAgentErrorCode is 32601.
For Google Dialogflow CX agent, the VirtualAgentProviderData JSON object may contain the following information. Note that this variable might be empty if the Widget exited through the Failed transition.
|ConversationId||ConversationId: Unique identifier for this conversation provided by Google|
|EndUserId||Unique identifier for the end user participant provided by Google|
|AgentHandoffParameters||Parameters included from the Dialogflow CX Agent if the Live Agent Handoff Transition was triggered|