Academy
Transformation

Transform AI Step

The Transform AI step combines text interaction with Dattos AI to make data preparation simpler and more effective. In a chat interface, you describe in natural language what you want to do with the data and see a real-time preview of the result. The step is located in the Preparation category of the toolbar.

With Transform AI you can, for example:

  • Add, delete, and change columns.

  • Reorder and rename columns.

  • Create expressions for calculations and operations.

  • Merge data from multiple sources.

Setting up Transform AI

  1. Add the step: drag Transform AI to the workspace and connect one or more steps to its data input.

  2. Type the command: write, in text, the action you want to perform. After each command, the change history appears in the chat, allowing you to track every change made to the report.

  3. Check the result: view the real-time preview in the Preview panel on the right, and see the original data in Connected tables.

  4. Apply transformations: click Apply Transformations to save your edits. This step is necessary for changes to be recorded.

Undo an action: to roll back a change, click the Restore to this action icon next to the performed actions box.

Best practices

  • Write clear and objective instructions: the more specific the description of what you want, the better the AI understands the objective.

  • Cite column names: mention the exact titles of the columns involved. This avoids ambiguity and helps build the correct transformation.

  • Review the result before applying: check the preview and ensure the output matches what you expected.

Data security

Dattos is committed to the security of your data. All communication with the AI only shares the column titles of the connected tables. Table data is never sent.

Usage example

Consider an input table with the columns Date, Entry, Value, and Observation, with entries from 02/05/2023 (values 100 and 200) and 03/05/2023 (values 150 and 300). When typing the command "Sum the value and group by date", the output provides two columns, Date and Value: 02/05/2023 with 300 and 03/05/2023 with 450. The example shows how the step sums the values and groups by date, simplifying the preparation.

  • Transform SQL Step

Was this helpful?