Semantic configuration for search indexes
You can improve the relevance of semantic search results by using semantic configurations. Semantic configurations provide precise control over search quality, eliminating the need for recreating the index.
TotalAgility provides a Default semantic configuration; however, you can add new semantic configurations when adding indexes to the AI knowledge base.
-
In the
New AI Knowledge Base configuration dialog box, select
for
Indexes.
The Add Index dialog box is displayed.
-
To add a new semantic configuration, select
for
Semantic configuration.
The Add Semantic configuration dialog box is displayed.
- Enter an ID for the configuration.
- By default, the ID appears as the Display name. Enter a different name. This name is used to identify the configuration when selecting it in the knowledge base activities or forms.
-
On the
Title field list, select an index field.
This title field is used by semantic search to represent the "title" of a document.
-
To add the content fields that represent the main searchable content:
- Select for Content fields.
-
On the
Display name list, select the system field and select
Add.
The content field is added to the table.
-
Add more fields as needed.
-
The content fields must be of type, "String" or "Collection of string".
-
You can add, remove, or re-order the content fields.
-
-
To add the keyword fields for use in semantic search:
- Select for Keyword fields.
-
On the
Display name list, select the system field and select
Add.
The keyword is added to the table.
-
Add more fields as needed.
-
The content fields must be of type "String" or "Collection of string".
-
You can add, remove, or re-order the keyword fields.
-
-
Select
OK .
The Add Index dialog box is displayed.
-
Select
OK and save the AI knowledge base configuration.
The semantic configuration is available for use in the AI discovery agent activity, Copilot insights, and the Generative AI chat control using the knowledge base.