[]
Datasets and data models that need to be made available for AI analysis need to have AI enabled on them and descriptions added to their fields so that the large language models (LLMs) understand the meaning of data accurately. This improves query accuracy and analysis efficiency. Through modeling, users can add descriptions, aliases, and enumeration values, allowing AI to better handle business scenarios during chat analysis.
To allow AI to understand the dataset content:
Open the dataset or model.
Navigate to the AI Settings page.
Set For AI conversation analysis to True.

In the portal’s Datasets or Data Models page, datasets/models available for AI chat analysis will display √ in the For AI conversation analysis column. If √ it is not displayed, the dataset/model cannot be used for AI chat analysis.

Add Description to Dataset
Provide an overall description of the dataset, including:
Business scenario.
Term definitions.
Business rules help the LLM associate understand context.

Add Aliases to Fields
For similar field names (e.g., RequiredDate and ShippedDate), add aliases:
RequiredDate → RequiredDeliveryDate.
ShippedDate → ActualShippedDateThis reduces ambiguity and improves query precision.

Add Field Descriptions and Enumeration Values to Fields
Add detailed descriptions for each field.
Define possible enumeration values (e.g., Active/Inactive for a status field) to assist AI in precise matching.


Add Description to Data Model
Provide an overall description of the data model, outlining:
Structure.
Business purpose.
Add Description to Entity
Add specific descriptions for each entity, explaining:
Its role.
Relationships in the model.
Add Field Descriptions and Enumeration Values to Fields
Add detailed descriptions for entity fields.
Define enumeration values to enhance semantic understanding and query accuracy.