Download
Community
Training
System Status
AU
Australia En
Canada En
Canada Fr
France Fr
Germany De
Italy It
Singapore En
Spain Es
Portugal Pt
United Kingdom En
United States En
MVP
Message
Video
Phone
Rooms & Connector
Webinar
Contact Center
AU
Australia En
Canada En
Canada Fr
France Fr
Germany De
Italy It
Singapore En
Spain Es
Portugal Pt
United Kingdom En
United States En
MVP
Message
Video
Phone
Rooms & Connector
Webinar
Contact Center
Download
Community
Training
System Status

Engage Digital | Configuring Agent Assist

Last updated on March 01, 2022

Table of contents

Configuring the Agent Assist extension

The extension becomes visible in the Engage Digital extension list once the feature flag is enabled by Support. 
 
You can then add and configure the extension.
  1. Go to Settings > Extensions via the left navigation bar.
  2. Click Add.
  3. Select Agent Assist. 
  4. Provide the Client ID and Client Secret.
  5. Change the location if you want to store data in another region.
  6. Click on Continue with Agent Assist.
  7. Sign in with your Google Cloud account. 
  8. Copy the Project ID and Refresh token.
  9. Click Save.
    10.  

Configuring the Agent Assist datasets

You can configure datasets for FAQs and Smart replies.

Configuring a FAQ dataset

To create a FAQ dataset, you need to prepare a CSV file with only two columns (question and answer).
  1. Go to Admin > Agent Tools > Agent Assist Datasets in Engage Digital.
  2. Click New Dataset > FAQ.
  3. Enter the name of the dataset. 
  4. Enter the language of the dataset (only English, Spanish, or French are available for now).
  5. Import the file.
  6. Click Save.

Configuring a Smart Reply dataset

Smart replies are created based on the conversations available in Engage Digital. The set of the conversations is prepared based on the conditions chosen in the dataset configuration.
  1. Go to Admin > Agent Tools > Agent Assist Datasets in Engage Digital.
  2. Click New Dataset > Smart reply.
  3. Enter the name of the dataset. 
  4. Enter the language of the dataset (only English, Spanish, or French are available for now).
  5. Choose Channels and Categories if you want to limit the conversations collected for the model.
  6. Click Save.
Wait for about 24 hours for the model to be fully configured and deployed.

Advanced configuration of the model

By default, all messages imported to the model appear in the Unreviewed messages tab. All messages available in Unreviewed messages and Allowed messages are available for future suggestions. Meaning that if you need to exclude some messages from the suggestions, you should block them. This configuration is available directly from the Google Agent Assist console.

Configuring an Agent Assist conversation profile

  1. Go to Admin > Agent Tools > Agent Assist.
  2. Click New Agent Assist.
  3. In the opened window, complete the profile options.
  4. Choose the datasets created previously (note: you can choose Smart reply, FAQ, or both).
  5. Click Save.
Once the item is saved, you can find it in the Google Agent Assist console at:

https://agentassist.cloud.google.com/projects/<id of GCP>/conversation-profiles

Agent Assist conversation profile options

  • Name: The name of the created item.
  • Language: The language of the created item (only English, Spanish or French are available for now).
  • Channels: The list of the channels where you want to see suggestions (messaging channels only).
  • Categories: The list of categories that the conversation has to get suggestions (only exact mapping).

Configuring the permissions needed in Engage Digital

Two permissions control access to the Agent Assist feature:
  • Manage Agent Assist: Grants permission to perform the Agent Assist configuration in Admin.
  • Use Agent Assist: Grants permission to see Agent Assist suggestions in Agent.
System Status
Legal
Cookie preferences
Privacy Notice
Contact us
© 1999-2022 RingCentral, Inc. All rights reserved.
Thanks!
We've sent you a link, please check your phone!
Please allow a full minute between phone number submissions.
There was an issue with SMS sending. Please try again. If the issue persists, please contact support.