Automations are actions that run automatically on all datasets added to a project.

​
What can you automate?

  • ⋈ Automatically import columns from another dataset.
  • ⚡ Automatically run evals.
  • ▶︎ Automatically run dynamic columns.
  • ⛓️ Automatically run flows.

​
When do automations run?

Automations will run when:

  • A dataset is moved to a project that has configured automations.
  • A dataset is created and assigned to a project that has configured automations.

​
How do you configure automations?

Automations live at the project level.

To configure automations, create a project from the Dataset List page and then click on the “Automations” button at the top right corner of the project.

A sidebar will appear on the right with the automations you can configure.

​
Automation Steps

Order Matters: Arrange steps logically - for example, import necessary columns before running evals that depend on them

​
Import Columns

Import specific columns from another dataset into your target dataset

  • Configure join conditions to match data between datasets
  • Select which columns you want to import

​
Run Evals

Automatically execute configured evaluations on new datasets

  • Select a dataset where you have previously run these evaluations.
  • The evaluation configuration will be copied from that dataset, and executed when a new dataset is added to the project.

​
Dynamic Columns

Automatically run dynamic columns on new datasets

  • Select a dataset where you have previously run dynamic columns.
  • The dynamic column configuration will be copied from that dataset, and executed when a new dataset is added to the project.