Documentation

    {{docApp.userSession.userId}}
    {{docApp.userSession.userName}}
    {{docApp.userSession.email}}
    Issuer : {{docApp.userSession.issuerInstance}}
    Expires : {{new Date(docApp.userSession.validUntil).toLocaleDateString()}}

{{docApp.title}}

{{docApp.description}}

How can we help?

{{docApp.searchError}}
{{product.name}}

Searching in {{docApp.searchFilterBySpecificBookTitle}}

{{docApp.searchResultFilteredItems.length}} results for: {{docApp.currentResultsSearchText}} in {{docApp.searchFilterBySpecificBookTitle}}
Search results have been limited. There are a total of {{docApp.searchResponse.totalResultsAvailable}} matches.

You have an odd number of " characters in your search terms - each one needs closing with a matching " character!

{{docApp.libraryHomeViewProduct.title || docApp.libraryHomeViewProduct.id}}

{{docApp.libraryHomeViewProduct.description}}

  1. {{book.title}}

{{group.title || group.id}}

{{group.description}}

  1. {{book.title}}

{{group.title}}

{{book.title}}
More...

Sentiment Analysis Configuration

Important

HAi features are currently in a closed beta. You can only access these features after you register. Contact your Hornbill Customer Success representative to join the program.

Before you begin

Ensure you have enabled HAi features. You can customize how much data the system uses and which timeline update types trigger a request summary. To manage these settings, navigate to the application settings for Service Manager.

Setting Description
generativeAi.sentimentAnalysis.availablePostTypes Determines the types of posts used to generate the request sentiment. Valid values include: Authorization, Customer, Email, Escalate, and update.
generativeAi.sentimentAnalysisAutoUpdates Enables or disables automatic request sentiment analysis when a customer provides an update.

Auto updates

When you enable Auto Updates, the system checks the sentiment of a request if the person posting or commenting is the customer associated with that request. This applies to the following scenarios:

  • Customer Portal: The customer adds a timeline post or comment.
  • Employee Portal: The customer adds a timeline post or comment.
  • Email Auto Responder: The system updates a request via an email sent by the customer.
  • Mailbox: A user applies an email to a request where the sender is the customer.

Data processing

When the system performs a sentiment analysis, it processes specific data from the active request.

Request fields

The following technical fields are processed:

  • h_itsm_requests.h_description
  • h_itsm_requests.h_summary
  • h_itsm_requests.h_fk_user_name

Timeline data

The system filters the request timeline using the following criteria: ["Customer","Email","update"]. For every post in the filtered timeline, the system sends the following data:

  • actorInfo.name
  • content

For any comments associated with these filtered posts, the system sends:

  • actorInfo.name
  • comment

Custom questions

The system passes the first 100 custom questions. It excludes file-upload and label types. The data is paired using the following format:

  • h_question
  • h_answer_value
  • Version {{docApp.book.version}}
  • Node {{docApp.node}} / {{docApp.build}}
In This Document