New AI Agent to Ingest External Files Containing Skill Data

You can now upload files containing skills data into the AI engine using a new agent called Skills Library Training, which can then be used by the Skills Library Enrichment agent to discover and suggest skills to add to your skills library. 

To use this agent, you need to include the skill data in a JSON file format and upload the file. Skill data can come from one or more of these object types: 

  • Requisitions 
  • Candidates
  • Applicants
  • Job
  • Skills
  • Person
  • Position
  • Catalog
  • Learning items

You can upload more than one JSON file to an agent process. Each file must contain data pertaining to a specific object type from above. 

When skills data from these object types is added to the AI object store, the AI engine analyzes this data and fine-tunes the AI skills model to provide better suggestions that are relevant to your business needs.

To configure and run the Skills Library Training agent:

  1. Go to My Client Groups > Learning and Development. In the HCM agents section, click Skills Library Training
  2. From the Add menu, select Skills Library Training
  3. Enter a name and description for the agent. 
  4. Upload a JSON file that contains the data you want to import. 
  5. Click Apply.  
  6. On the Learning and Development Agents page, select the process you just created and click Run Agent.

JSON File Specifications

  • Each JSON file should be less than or equal to 512 MB. The file name must be in this format: Source_Objecttype_yyyyMMdd_HHmmss.json.
  • For the object type in the file name, you need to choose from one of these: requisitions, candidates, applicants, job, person, position, catalog, skills, or learning items. For example, a valid file name would be: Oracle_person_20240817_123045.json.
  • The file name must be unique and shouldn’t have been used in any other existing process, including completed processes.
  • The metadata definition in the JSON file must contain these fields: source, object_type, version, and isoCreationDate.

Editing a Completed Process

  • After the process completes, you can edit it by changing the name and description or adding more attachments. If you add attachments and run the process again, the previously uploaded attachments won’t be processed. Only the new attachments will be processed.
  • You can delete a completed process. This won’t retract the data that’s already added to the AI object store.

This feature offers an easy and flexible way to use external data from skill sources such as candidates, requisitions, jobs, and so on, to train the AI skills library.

Steps to Enable

You don't need to do anything to enable this feature.

Access Requirements

You need to have the Manage Bulk Skills Library Training privilege (WLF_MANAGE_BULK_SKILLS_LIBRARY_TRAINING) to run this agent.