LogoLogo
  • Go to Docs Center
  • Meet Plum Fuse
  • Tutorial
    • Basics
    • Deployment
    • Building A Complex Application
  • How to...
    • Send SMS messages using the REST module
    • Use the SMS Module
    • Upload an Audio Recording to Storage
    • Use Plum's Transcription API
  • Fuse Navigation
  • Application Manager
    • New Application
    • Managing Applications
    • Folders
  • Template Center
  • Deployments
    • Deployments Table
    • Creating Deployments
    • Updating Deployments
    • Deleting Deployments
    • Outbound Deployments
  • Logs
  • Reports
  • My Account
    • Global Options
    • Admin Options
  • Users & Sharing
    • Single Sign On
  • Application Editor
    • Application Editor Overview
    • Settings
      • Language Settings
        • Text-To-Speech (TTS) Options
      • User Input Settings
      • Connection Settings
      • Post-Call Webservice
      • JavaScript Libraries
    • Audio Manager
      • Languages
      • Prompt Table
      • Audio Formats
      • Bulk Uploading Audio Prompts
    • Modules
  • Modules
  • Call-Flow
    • Prompt
    • Transfer
    • Go To App
    • Go To Page
    • Compare Variable
    • Counter
    • Exit
    • Call Log
    • Label
    • Subdialog
  • Input
    • Address
    • Yes/No
    • Custom Field
    • Date/Time
    • Number
    • Digits
    • Language
    • Menu
    • Dynamic Menu
    • Name
    • Record
  • Data
    • Email
    • SMS
    • Set Variable
    • Call Recording
    • Evaluate JS
    • REST
    • SOAP
  • Integration
    • Setting up authentication
    • Dialogflow
    • DynamoDB
    • AWS S3
  • Variables
    • Shadow Variables
  • Module Settings
  • Key Fuse Info
  • 📞Outbound
    • Queuing Calls with CSV
    • View Pending Calls
    • Viewing Completed Calls
    • Detecting Voicemail
    • Outbound Parameters
    • Outbound FAQs
  • 🔒Data Security
    • 'Private' Mode
    • Managing Secure Phone Numbers
    • Sensitive Data Types
  • 🔑APIs
    • Authentication
    • Outbound Calls
      • Queue Call
      • Queue Multiple Calls
      • Get Outbound Call Status
      • Cancel Outbound Calls
    • Logs
      • Get Call Logs
      • Get Call Logs With Details
      • Get A Detailed Call Log
  • 🗒️Release Notes
Powered by GitBook
On this page
  • Function
  • Prerequisites
  • Input Fields
  • Shadow Variables (Dialogflow only)
  • Module Settings
  • Hold Music
  • Timeout Length
  • Private
  • Barge-in
  • Show Custom Errors
  1. Integration

Dialogflow

PreviousSetting up authenticationNextDynamoDB

Last updated 2 years ago

IMPORTANT: This is an advanced module. Some setup is required before use.

Function

This module uses the supported third-party service Google Dialogflow. The module allows you to integrate a Dialogflow voice agent into your application to interact with callers.

You can select and use any Dialogflow integration that has been added to your Fuse account.

NOTE: Integrations are shared with all users in a Fuse account. You can always check existing integrations by going to the My Account page and viewing the Integrations tab.

Prerequisites

To use this module, you must have access to Google Dialogflow. You must also have a working Dialogflow voice agent added as an integration to your Fuse account.

Complete the following in order:

In Google Dialogflow:

In Fuse:

Input Fields

As shown in Figure 2, the following fields are available:

  1. Select Integration: Select one of your Dialogflow integrations.

  2. Select Environment: Select a Dialogflow environment. List options are retrieved from Google Dialogflow based on the selected integration.

  3. Prompt field: Enter a message that you want your voice agent to read when it starts. This field accepts strings and/or variables.

Shadow Variables (Dialogflow only)

The Dialogflow module also supports the following shadow variables:

Shadow variable
Returns

name$.query

string End-user input received by Dialogflow.

name$.intent

string Intent matched to the received end-user input.

name$.message

string Text of the response spoken or displayed to the end-user based on the matched intent.

name$.entities

object Entity types associated with the matched intent.

name$.confidence

integer Confidence level for the name field.

name$.allRequiredParamsPresent

boolean Returns true when:

  • All required parameter values have been collected

OR

  • The matched intent contains no required parameters.

Returns false when:

  • The matched intent has required parameters

AND

  • One or more required parameter values have not been collected.

Module Settings

This allows users to set the maximum time, in seconds, that elapse before returning a timeout error. This applies to modules that fetch outside scripts, like the SOAP, REST, or subdialog modules. The default value for Timeout Length is 30 seconds. Acceptable number values range from 1 to 120.

Get access to Google Dialogflow. ()

Build your own working, custom voice agent. () (Optional) You may also use a prebuilt voice agent instead. ()

Set up a service account and service account key for your agent. ()

Admin-only task: .

Once finished, the Dialogflow module can be used successfully.

See for more information on this variable type in general.

Enable this setting to play audio while an end-user is on hold. To upload a file for hold music, go to Application Settings > Connection Settings > > Webservice fetch audio.

This setting controls the logging function of a module. Enabling the 'Private' setting instructs the module to not record, report, or retain the information input to that module for reporting or any other purposes. When enabled any information entered into a module during a call will be lost immediately when the call terminates. The 'Private' setting is critical for businesses that need to maintain PCI-DSS or HIPAA compliance. The module icon, in the upper left-hand corner, becomes grayed-out when this setting is enabled. See more details .

This setting allows callers to interrupt a prompt before it finishes playing. When enabled, DTMF input interrupts the prompt and progresses the call forward in the call-flow. If speech recognition is enabled on the ensuing module in the call-flow, then end-users can also interrupt the prompt with a spoken utterance, too. Disabling barge-in forces callers to listen to the entire prompt. A barge-in enabled module will have a dashed line on top of the text box. See example .

Enabling this setting overrides the default, global error options set in the Application Settings > . This allows users to establish custom error handling in order to act on errors in a specific way in that module. Instead of progressing to the next module in the call-flow, custom error handling allows users to re-prompt the same module, to provide a custom error message, to re-direct the call based on the error, or any other desired behavior. Adding multiple errors () to a module functions behave the same way as a counter. The first error follows the path for the first error listed, if a second error occurs in the same module it follows the second listed error, and so on until all errors are exhausted or an error directs the end-user away from that module. No Input occurs when the caller does not provide an input based on the timeout settings. This is based on the “Initial input timeout” in . No Match occurs when the caller input does not match the module's criteria for the input module.

Dialogflow website
Google Documentation
Google Documentation
Shadow Variables
User Input Settings
User Input Settings
Webservices
Add your voice agent as an integration in the My Account page
Guide
Hold Music
Timeout Length
Private
Barge-in
Show Custom Errors
here
here
Figure 1: The Dialogflow module in use.
Figure 2: Dialogflow module fields.
Figure 3: Module settings for Dialogflow.