Create a new Knowledge Source

Navigate to the list of Knowledge Sources and click the New button.

A new Knowledge Source pop-up will appear.

Fill in the Knowledge Source fields:

• Name
• Description
• System name. System name is used to identify the Knowledge Source e.g. when configuring RAG Tools, so it is strongly recommended not to change it once a Knowledge Source has been created.
• Source settings. Indicates the location where information available for the RAG is stored. These settings cannot be changed once a Knowledge Source has been created. Follow these instructions to fill in source settings:

If you are on Ideaport demo environment:

• Source type: Sharepoint
• Sharepoint URL: URL of Sharepoint subsite that stores your content, e.g. https://ideaport.sharepoint.com/sites/GenAI/Example
  Link to subsite will be provided by Ideaport. 

If you are on your own environment:

• Source type: depending on your setup, this could be Sharepoint, Confluence, Salesforce, or RightNow. Contact the admin responsible for Magnet AI setup to find out what types of sources are available.
• Depending on Source type: source parameters, e.g. source URL.

• Advanced: Max chunk size. When content is embedded into vector database to become searchable by the RAG Tool, it is split into chunks (pieces) for best performance. Max chunk size defines the maximum number of characters in a text chunk. The default value is 1000 characters, but it might be useful to experiment with different chunk sizes, making pieces of embedded content longer or shorter.

Once you have filled in all the Knowledge Source parameters, click Save & Sync. This will embed your content into vector store and make it available for RAG Tools.