Call Dialogflow Bot action
Integrations > Integrations > View permission assigned to any user role
Use the Call Dialogflow Bot action to run self-service applications within a flow before or instead of routing a customer to an agent. Use this action in inbound and outbound call flows, and chat and message flows.
- This action is available in call, chat, and message flows.
- Google Dialogflow bots that use the Genesys Virtual Agent Services powered by Google CCAI integration in Genesys Cloud are not PCI compliant. Best practice recommends that you do not use them in Architect secure call flows.
- To enable the Call Dialogflow Bot action in Architect, you must first obtain the Google Dialogflow integration from the Genesys AppFoundry. Then, configure it in Genesys Cloud and create a Google Dialogflow bot in the Google console.
- You must have at least one active Google Dialogflow integration enabled. For more information, see About the Google Dialogflow integration.
- Changes you make to a Google Dialogflow bot in the Google console can take a few minutes to appear in Architect.
- To enable barge-in functionality for a Google Dialogflow bot, see Enable barge-in for Google Dialogflow bots.
- Administrators can use any text-to-speech provider that is available in Architect flows for use with their installed bot integrations. For more information, see About text-to-speech (TTS) integrations.
- For further assistance on configuring Google Dialogflow, see About the Google Dialogflow integration and Google Dialogflow FAQs.
|Name field||Type a distinctive name for the action. The label you enter here becomes the action’s name displayed in the task sequence.|
|Agent Input Text||
The string variable text to send to the Dialogflow Agent. Typically, you only use this field when you chain bot actions together.
Note: Agent Input Text is only valid for digital. You can control the initial routing by using bot variables in Architect with conditions in Nuance Mix bot. When you implement this scenario, set the bot input text to “”.
|Follow-up Response Timeout||
If the bot must ask for follow-up information, use this field to set how long the bot waits for the user to send a reply. The system treats a NOT_SET, negative, or 0-second duration as immediate.
The specific Dialogflow bot, also called agent, that you want to use. This version is the agent bot that you create in the Google console.
Select the Google Dialogflow environment that you want the flow to use at runtime.
Note: A draft environment is always available for a Dialogflow agent. If you did not create any environments for the selected Dialogflow agent, then the only environment available is Draft. Otherwise, a list of published environments, in addition to Draft, appears.
Environments provide a way for a flow to select which version of a Google Dialogflow agent that Architect calls at flow runtime. For more information about using Google Dialogflow environments in Architect, see Use Google Dialogflow environments in Architect. For more information about Google Dialogflow environments, see Google Cloud Versions and Environments.
|Initial Event Name||
The name of the Dialogflow agent event to trigger when the interaction begins.
Information that you can pass into and retrieve from the bot. Enter the name of the variable and a value to assign. You can add multiple variables, and you can reorder them.
In digital flows, a Dialogflow Agent only uses input session variables when the flow supplies a non-blank initial event name. Use inputs to pass data to a Dialogflow agent. For each value you send, give the entry a name and an expression that resolves to a string for its value. Then when the action runs, all expressions evaluate to produce a string that passes to Dialogflow as event parameters. You can pre-populate slots by choosing the name of what passes as the name of your slot, or pass more data that you process in a Dialogflow agent’s webhook.
Use outputs to extract data from Dialogflow. Create a Dialogflow output context named “genesys-output” with name-value parameters. Then in Architect’s Output section, list the names of the parameters you want along with the Architect variables in which to store the matching values.
Note: Key names are case-sensitive.
The options that appear in this section reflect the logic configured in the Draft environment of the Dialogflow bot. Each intent represents an action to perform. For example, order a pizza or order flowers. Each intent path matches an output path for the Call Dialogflow Bot action in the task editor.
Below each intent, Architect displays the configured Dialogflow bot slots, or any data the user must provide to fulfill the intent. Using the Pizza Order intent as an example, the slots can include crust type, ingredients, cheese amount, and cutting style. If the user chooses the Order Pizza path, then the Dialogflow bot provides the slots associated with that path. Regardless of the environment you select, the Call Dialogflow Bot action always shows the intents and slots from the Draft of the Dialogflow Bot.
Note: To ensure that the intent appears as an output in Architect, when you create an intent in Google Dialogflow, make sure to designate it as an end of conversation intent. Otherwise, the intent does not appear here.
The confidence of the detected Dialogflow end of conversation event. If the flow detects the intent, the value ranges from 0.0 to 1.0. If the flow does not detect an intent, the value of the variable that binds to this setting remains untouched.
Select the variables that you want to assign to the failure data values returned by the Dialogflow bot.
Note: To expand or collapse the list of variables, click the arrow next to Failure Outputs.
Configure intent and failure paths
The Google Dialogflow bot’s Draft dictates these paths.
|Failure output path||
This path indicates that an error occurred while running the action or there was a problem processing the results.
Note: If the network experiences connectivity issues, the action automatically takes this failure path.
Drag the appropriate action below the Failure path and direct the route you want the interaction to take. For example, a transfer action to send the caller to an agent for assistance.