Augment Data with Table AI Assist in Autonomous AI Database

Use Table AI Assist in Data Studio to prepare, enrich, and correct tables in your Autonomous AI Database schema using natural language prompts. Table AI Assist can generate SQL to add, replace, remove, or rename columns, and it can help derive new data, such as reporting periods from date columns or distances from geographic coordinates. The generated SQL can be used to create or replace a view, create a new table, alter the source table, or create a query, depending on how you select to apply the generated recipe.

This topic provides an entry point for understanding when and how to use Table AI Assist with Autonomous AI Database. For complete reference, see the Data Studio documentation.

When to use Table AI Assist

Use Table AI Assist when you want to prepare data for analysis without manually writing every SQL expression.

For example, use it to:

• Add derived columns, such as day, month, quarter, or year from a date column.
• Add calculated columns, such as distance between two coordinate columns.
• Replace values in a column using generated SQL.
• Remove or rename columns as part of a data-preparation workflow.
• Create a reusable recipe that captures a sequence of transformation steps.
• Preview generated results before creating a table, creating or replacing a view, altering the source table, or producing a query.

Table AI Assist is useful when you know the data-preparation result you want, but you prefer to describe the change in natural language and review the SQL generated by the AI service before applying it.

How Table AI Assist works with Autonomous AI Database

Table AI Assist runs in Database Actions for your Autonomous AI Database. Database Actions is bundled with each Autonomous AI Database instance and provides development tools, data tools, administration, and monitoring features. Data Studio is one of the feature areas available from Database Actions.

When you use Table AI Assist, the tool works with tables in your database schema and uses your configured AI profile to generate SQL from natural language prompts. You can review the generated SQL, preview the results, and then decide whether to save the work as a recipe, create a view, create a table, alter the source table, or use the generated query.

For non-ADMIN users, access to Data Studio tools requires the appropriate database role. In Autonomous AI Database, the required role for Data Studio tools is DWROLE. If you do not see Data Studio or the Table AI Assist entry point in Database Actions, ask an administrator to verify your Database Actions access and Data Studio role grants.

See The Table AI Assist Tool for an overview, examples, and prerequisite link to Data Studio Settings.

Before you begin

Before using Table AI Assist, configure an AI profile in Data Studio Settings. The AI profile controls how natural language prompts are translated into SQL statements. Data Studio Settings provides an AI Profile tab where you can create, select, validate, and manage AI profiles.

To use AI services in Data Studio, you must have access to a supported AI service account and access to the DBMS_CLOUD_AI package. Data Studio can create AI profiles in a guided manner, including profiles for natural-language-to-SQL use cases.

See Data Studio Settings for help with configuration for AI profiles and related Data Studio settings.

Open Table AI Assist

To open Table AI Assist:

1. Open Database Actions for your Autonomous AI Database.
2. Select Data Studio.
3. Select Table AI Assist.
4. Select a table in your schema.
5. Create or open a recipe, enter a natural language prompt, review the generated SQL and preview, and then select how to apply or save the result.

For complete task steps, use the SQL Developer Web and Database Actions documentation links in the following section.

What you can do from Table AI Assist

• Prepare data with recipes: Build a recipe that is made up of one or more steps. Each step can add, replace, remove, or rename columns without immediately changing the base table. Recipes can be saved, exported as JSON, imported, and reused.

  See Prepare Data Using Table AI Assist Tool for step-by-step procedures for creating recipes, generating SQL, and previewing results.

• Create a view or table, or alter the source table: Apply recipe steps by creating or replacing a view, creating a new table, or altering the source table. See Prepare Data Using Table AI Assist Tool for more information.

  See Prepare Data Using Table AI Assist Tool for step-by-step procedures for creating target tables or views, exporting recipes, and importing recipes.

• Generate SQL from prompts: Ask Table AI Assist to generate SQL from a natural language prompt, then preview the resulting data before saving or applying the step. See Prepare Data Using Table AI Assist Tool for more information.

  See Prepare Data Using Table AI Assist Tool for step-by-step procedures editing recipes.

• Use conversational refinement: Continue a conversation with Table AI Assist so follow-up prompts can use previous context instead of being treated as isolated requests.

  See Conversational AI for the Table AI Assist Tool to learn how follow-up prompts use conversational context while generating SQL and previewing new columns.

• View table entity details: Use table actions to view details, gather statistics, share, register to Cloud Link, publish to Marketplace, create analytic views, query, export, edit, or drop a table, depending on available permissions and context.

  See View Table Entity Details to learn about the table actions available from the Table AI Assist page.