Starburst Agent#

The Starburst Agent is a chatbot that helps you analyze data by converting natural language questions into SQL queries and analyzing their results. Starburst Agent can generate and execute SQL queries and provide metadata about available datasets.

Note

Starburst Enterprise Agent is available as a private preview. Contact your Starburst account team for more information.

Requirements#

To use Starburst Agent, you need:

  • A valid AI_WORKFLOWS license.

  • Access to at least one configured language AI model.

Considerations#

  • Starburst Agent sessions are stored in coordinator memory and never persisted. Each session supports a single SQL query execution. The Agent can answer as many questions about the data produced by the query as desired. To ask a question that requires a new query, you must replay reset the session. The Agent can respond to both data-related and metadata-related questions. For example, you may ask questions such as Which columns are related to customer behavior?.

  • Starburst Agent is conversational. You may refer to previous questions and answers. For example, after asking which columns are related to customer behavior, you may follow up with, Show trends in customer spend over the last two years, grouping by the customer behavior dimensions identified in your previous answer.

  • Data product enrichment affects Starburst Agent’s response quality. For example, the more detailed a data product’s metadata is, the higher quality the agent’s response may be.

Configuration#

To configure the Starburst Agent, add the following property to your coordinator configuration file:

starburst.agent.enabled=true

Using Starburst Agent#

To open the chat dialog and begin a session with Starburst Agent:

  1. Navigate to the Data products tab in the Starburst Enterprise web UI.

  2. Select an existing data product.

  3. Click the Starburst Agent’s icon at the bottom-right of the screen.

  4. In the chat interface:

    • Use the left drop-down menu to select a persona.

    • Use the right drop-down menu to select an AI model. If only one model is configured, it is preselected.

  5. Enter a question or prompt in the text area.

  6. Press Enter or click the send submit button.

Edit a query#

When Starburst Agent decides a question can be answered by generating a query, it returns the query to the user for review.

To edit a query, click the Edit button. Once you have finished editing your query, click Save. To execute the query without editing, click the Continue button. This sets the query for the session.

Queries are validated. If a query fails validation, the Agent attempts to correct it up to two times by default, this is configurable. If it cannot be corrected, the error message is displayed in red text. A query cannot be executed unless it passes validation. Line and column numbers in validation error messages may be shifted due to formatting. Validation is not guaranteed to catch all issues. Errors are still possible at runtime due to data issues and invalid function arguments.

Add usage example#

When Starburst Agent decides a question can be answered by generating a query, it returns the query to the user for review. Starburst does not sample data, however, you can add sample data to the query.

To add sample data to your click the Add usage example button. Provide code snippets and use-case examples to help others explore the data product.

For example, if the description of DESTINATION_COUNTRY is “Country to which a shipment is being sent to”, the user may want to edit this to say “Country to which a shipment is being sent to. Sample values include ‘US’, ‘JP’, and ‘IN’”.

This would help the LLM to understand that it’s a two letter country code rather than it using values such as US, USA, United States, or United States of America.

The following table describes the icons used in the chat dialog:

Starburst Agent chat dialog icons#

Icon

Description

replay

Reset session.
content_copy

Copy the Agent’s response to clipboard.

download_2

Download the Agent’s response.

send

Submit question.

close

Minimize the Agent dialog.

Personas#

The Starburst Agent supports three personas. Each persona tailors its responses to suit different user roles and goals. The following sections describe the Executive, Analyst, and Data engineer personas.

Executive#

Provides high-level summaries tailored to executives and decision-makers.

  • Focuses on business insights and trends

  • Omits technical detail unless explicitly requested

  • Presents concise bullet points for quick understanding

Analyst#

Offers detailed analytical summaries suitable for analysts and data scientists.

  • Includes statistical analysis and relationships in the data

  • Adds contextual information and potential implications

  • May include suggestions for further exploration

Data engineer#

Provides technical summaries tailored to engineers.

  • Focuses on structure, data quality, and metadata

  • Includes schema details, cardinality, and patterns

  • Highlights potential data issues or anomalies