Skip to main content

Learning status

The status of the file can also be changed to one of the following options:

  • Suspended
    • When a file is set to suspended, the associated agents to the topic of the document will be able to consume the content of the file using the last snapshot taken before it was suspended.
  • Published
    • Whenever the document changes content or its topics are updated, the changes will be reflected in the agents answers upon the completion of the learning process
  • Draft
    • Available only if the file is created without publishing it. The agent will not be able to consume the content of that file until it is set to published. Once set to published the draft status will be removed from the available options.
  • Archived
    • It allows the user to mark a document as no longer part of the knowledge of any agent without deleting the file. To restore the file learning previously generated, the user can set the status back to published.

When changing the status of a file, a prompt in the top right of the page will provide details about the change.

As the status is changed, the learning messages will change:

  • 1.Learning not started: This is the default state for an uploaded document that has not begun the learning process.
  • 2.Learning pending: ToothFairyAI is getting ready to start the learning process
  • 3.Learning in progress - %: Shows the progress of the learning process
  • 4.Learning completed: The document is not ready for one of the agents to use the retrieve information.

The time it takes for the AI agents to learn about a document depends on the size of the document.

The user can keep track of the progress based on the completion percentage displayed. The process is done in the background and does not affect the performance of the platform nor it requires the user to stay on the page.


It is recommended to limit the changes to the documents while the agents are in use for production purposes.

The changes made to the documentation will be reflected in the agents' answers only after the learning process is completed.