Auto generate rule expression for pass conditions using AI

The auto rule expression generation feature lets you set advanced pass conditions in Quality Rules and Pipelines by describing their logic in plain English. This AI-powered capability interprets user intent and dynamically generates precise expressions, streamlining the rule creation process and reducing manual effort.

This feature is available in the Basic and Advanced tab of the Pass conditions, Add Evaluated Rows Filter section in Create Rule page, and Condition tab in Step Configuration of a Data Quality pipeline and is powered by the AI Assist button. The button is designed to support expression generation.

Benefits

• Simplified expression creation: You can describe your logic in natural language, and the application converts it into a valid expression using field names and data types.
• Reduces manual effort: You don't need to write complex scripts as the AI handles the conversion, saving time and minimizing mistakes
• Consistent and accurate output: Expressions are generated using standardized grammar and metadata, ensuring uniformity across rules and pipelines.
  Tip: To ensure your expression is generated accurately, follow the instructions below:

  • Use valid field names from your dataset. The application uses these names and their associated data types to construct accurate expressions.
  • Avoid unsupported keywords, such as Value alias, which can prevent the application from generating a valid expression.
  • Verify prompt accuracy. If your prompt references the fields that do not exist, expression generation may fail.
  • Refine unclear prompts. When needed, revise your input using valid field names and supported syntax.

To generate the expressions in the pass conditions:

1. On the main navigation menu, click Quality and navigate to Rules.
2. On the Rules page, click on the Create Rule button
3. Select the target assets that have to be evaluated by the rule and proceed to define the pass conditions.
4. Click the AI Assist button present beside the pass conditions section to open the expression generation panel.
5. In the description box, enter your logic in plain English. Ensure the details of the relevant datasets or fields are added to the logic.
6. Click Generate. The application will automatically generate an expression for the pass condition based on the inputs.
   Tip: If the generated pass condition doesn't produce the desired outcome, you may revise the prompt and select Generate to create a new pass condition based on the specified parameters. Once the expression meets your requirements, you can either copy the generated expression or select Apply to implement the pass condition within the rule.
7. Click Create to save the rule with the generated expression and other details.