3.1 Understanding AI Enrichment

AI Enrichment enables you to enhance your database by systematically adding, refining, and optimizing schema, table, and column metadata.

With this feature, you can assign clear, meaningful labels and comprehensive descriptions to database objects including tables and columns. By using annotations to declare the intent of your data, you can make your database more understandable and AI-ready for LLMs and natural language querying. This enrichment enables LLMs to generate more accurate, context-aware SQL queries.

3.1.1 Enabling AI Enrichment for a Schema

To view enrichment opportunities for your schema and take action, you must first enable the AI enrichment feature.

This is a one-time process that sets up your schema by creating the necessary objects to store the enrichment metadata.

To enable the feature:

1. In the Connections panel, expand your database connection tree, and click the AI Enrichment folder.
   
   

   

   
   
2. In the dialog box that appears, click Yes to confirm the creation of AI enrichment objects.
   
   

   

   
   

   The AI enrichment feature is now available for your schema.

   Note:

   Do not use AI Enrichment on SYS or SYSTEM schemas, as the creation of AI Enrichment objects in a SYS-level schema may fail due to potential conflicts during installation.

3.1.2 Required User Privileges for AI Enrichment

To successfully install and use the AI Enrichment feature, the following database privileges are required:

• CREATE VIEW
• CREATE TABLE
• CREATE SEQUENCE
• CREATE PROCEDURE

Note:

The user must also have sufficient quota in the tablespace.

3.1.3 Introduction to the AI Enrichment Dashboard

The AI Enrichment dashboard provides a consolidated view of database schema, allowing you to monitor and manage schema enrichment activities in one place. With this dashboard, you can easily track the progress of annotations and enrichment across tables and columns, helping ensure completeness throughout your schema.

To open the AI Enrichment dashboard:

In the Connections panel, expand your database connection and click + on the AI Enrichment node. Alternatively, right-click the AI Enrichment node and select AI Enrichment Dashboard.

The AI Enrichment Dashboard is divided into following key sections:

• About this schema: Allows you to enter the high-level description for the schema. For more information, see High-Level Schema Description.
• Table groups: Allows you to logically group related tables and displays the existing groups in the schema. For example, all tables related to Human Resources can be grouped into an HR table group. For more information, see Managing Table Groups.
• Database Enrichment Percentage: Allows you to view the percentage completion of Al Enrichment in your schema. If none of the schema objects (tables or columns) are annotated, the enrichment percentage will be 0%. As more tables and columns within the schema are enriched, this percentage increases accordingly. This metric helps you track progress and measure how extensively enrichment has been applied within a given schema.
• Suggestions detected: Displays helpful hints and recommended next steps to guide users in maximizing the effectiveness of schema enrichment.
• Export to JSON: Allows you to export selected schema, table group, table, and column enrichment data to a JSON file.

3.1.4 High-Level Schema Description

The first step in schema enrichment is to add a high-level description for your schema.

You can use this section to explain the purpose of your schema and clearly communicate its role and structure within your database. Adding this description helps users and tools quickly understand the business context, objectives, and key entities represented by the schema.

In the AI Enrichment dashboard, add a clear and concise description in the About this schema section. For example, for a Human Resources (HR) schema, you can enter a description as follows:

This schema manages core HR data including employees, departments, roles, payroll, and benefits to support workforce management and reporting.

3.1.5 Exporting AI Enrichment Data to JSON

Use Export to JSON to export selected AI enrichment metadata for a schema, table groups, tables, and columns. Exporting annotations to a JSON file enables you to version-control, programmatically analyze, and share your schema annotations.

The exported annotation metadata includes the label, value, identifier type, and identifier name. When applicable, it also includes the table name and column name.

Perform the following steps to export AI enrichment data:

1. In the AI Enrichment dashboard, locate the Actions section.
2. click Export to JSON.
   The Export to JSON slider appears.
   

   
   Description of the illustration ai_enrichment_export_to_json.png

   
   
3. Select the enrichment data to export.
   You can use the following options:

   • Select all: Selects all available enrichment metadata.
   • Search: Filters the available schema objects.
   • All: Shows schema, table groups, and tables.
   • Table groups: Shows only table groups.
   • Tables: Shows only tables.
   • Schema description: Exports the schema description.
4. Expand Table groups or Tables to select specific child objects.

   Table groups: You can expand a table group to view its tables. Expand a table to view its columns. Select the check box for a table group, table, or column to include it in the export.

   Tables: You can select individual tables or expand a table to select specific columns. By default, a limited number of tables is displayed. Turn on Show all tables to display the full table list, including tables that are part of table groups.

5. Click Export Selected to export the selected AI enrichment data.
6. In the file save dialog, select a location and save the JSON file.

   The default file name uses the schema name, for example, HR_AIEnrichedData.json