Manage Bot Configuration

This article deals with the following topics:

The Bot Configuration page provides the bot details and includes two tabs – Bot Information and Follow-up. You can set variables that are specific to a bot in a single workspace. You can manage bot configurations such as the Natural Language Processing (NLP) threshold, Welcome Skills ordering, Abort Skill, Post Fallback, Threshold Validation, Channels, Follow-Up Skills, and Satisfaction Survey.

To access Bot Configuration, navigate to the Bot Dashboard, click the More icon and select Bot Configuration

Bot Information

The Bot Information tab includes the following three sections:

  • Bot Details
  • Natural Language Processing Settings
  • General Settings

Once you make changes to any of the values on this tab, click Save to save the changes, else click Cancel to navigate to the Bot Dashboard

Bot Details

The Bot Details section includes the following bot details:

  • Bot Name: Specifies the bot name. This is used to identify the bot easily in Skills Builder.
  • Bot Language: Displays the language that the bot can communicate with the user. Currently, a bot supports only one language.
  • Last Updated By: Displays the user name who most recently modified the bot.
  • Last Updated On: Displays the date and time of the most recent update to the bot.
  • Bot External ID: This is the unique identifier assigned to your bot. This ID is used to integrate with external applications or channels that require to communicate with the bot.

Natural Language Processing Settings

In the Natural Language Processing Settings section, you can automate the training with user responses and spelling correction for your bot, so that you can ensure the bot's NLP processing responds to your needs. To view the status of Natural Language Processing Settings, go to Bot dashboard and click  > Bot Configuration > Bot Information > Natural Language Processing Settings

Set the following NLP values for your bot:

  • Automate Training with User Responses: Luma prompts the user with a validation message when their phrase does not have a strong NLP match with a skill. The message asks the user if this is the skill they were looking for with Yes/No response options. If the user responds with Yes and automated training is enabled, the phrase will be automatically added to the skill. Under Suggested Training, administrators can review the phrases that were used and assign or un-assign them from skills. Suggested Training includes the following scenarios:

    Scenario 1: Automate Training with user responses is set to Active

    • On the Bot dashboard, click  > Bot Configuration > Bot Information > Natural Language Processing Settings > Automate Training with User Responses > Active.

    • While executing the skills in Luma, if a user enters a phrase that matches with any of the suggested skills available in Luma, then Luma displays the closest matching skill and prompts for confirmation (with Yes/No options) whether that is the required skill. Luma analyzes the sentence or phrase entered by the user, and based on the Natural Language Processing (NLP) score, suggests the skill which has the highest NLP score or threshold value. To learn more about how the suggested training option works, refer /wiki/spaces/CloudITSM/pages/880803841.

    Scenario 2: Automate Training with user response set as Inactive

    • On the Bot dashboard, click  > Bot Configuration > Bot Information > Natural Language Processing Settings > Automate Training with User Responses > Inactive.

    • While executing the skills in Luma, if a user enters a phrase that matches with any of the suggested skills available in Luma, then Luma displays the closest matching skill and prompts for confirmation (with Yes/No options) whether that is the required skill. To learn more about how the suggested training option works, refer /wiki/spaces/CloudITSM/pages/880803841.
  • Automatic Spelling Correction: If this is Active, misspelled words entered by the user are auto-corrected before passing to the NLP engine for evaluation. When active, the troubleshooting and debug logs will show what the user entered and what the NLP received so you can see how spellings are corrected. To activate this feature please reach out to the Serviceaide support team and raise a support ticket. Ask specifically to Activate/InActivate the Automatic Spelling Correction.

General Settings

Guest User Welcome Message: This is the welcome message that will be shown to the guest or non-registered users when they first start a conversation with the bot.

Common Error Message: This is the error message that will appear to the End Users and in the processing logs when Luma receives a random phrase that does not match any skill or encounters an error. 

The messages are configurable and can be updated as per your organization's needs. 

Welcome Skills

On the Welcome Skills tab, you can set and view each role's order of skills displayed within the Welcome Skill. To learn more, see the Define Welcome Skills page.

Abort Skill

On the Abort Skill tab, you can how users exit a skill before fulfillment occurs and how the bot will respond afterwards. To learn more, see the Manage Aborted Skills page.

Post Fallback

On the Post Fallback tab, you can specify how your bot responds after fallback occursTo learn more, see the Manage Post Fallback page.

Threshold Validation

On the Threshold Validation tab, you can customize how bot responds after the configured skill matching threshold value is reached. To learn more, see the Manage Threshold Validation page.

Channels

On the Channels tab, you can configure the channels or messaging platforms where your bot will be available. To learn more, see the Configure Chat Channels page.

Follow-up

On the Follow-up tab, you can configure the Luma follow-up processes. To learn more, see the Automated Follow-up Skill page.

Satisfaction Survey

The Satisfaction Survey tab helps you to customize a satisfaction survey, which can be triggered directly in the chat conversation. To learn more, see the Measure Customer Satisfaction page.