Skip to content

Generate Test Cases from Atlas

2 min read

Send an item or a journey from your map to the AI test case generator, so the tests it writes fit where that part sits in your app.


Atlas doesn’t write test cases itself. It opens the AI test case generator with two things filled in:

  • The requirements linked to the item, already selected.
  • A description of where the item sits on the map — what it’s nested in, which screens lead to it, where it leads, and what it depends on.

Told only “test the payment form”, the generator writes generic payment tests. Told “the payment form sits under Checkout, is reached from the cart and depends on the payment service”, it writes tests that cover those connections.


  1. Find the item on the map. Then either open the ⋮ menu on its row in the Tree view, or right-click it in the Map view, and choose Generate test cases… You can also select the item and use the sparkle button at the bottom of the inspector.

  2. The AI test case generator opens with the item’s requirements selected and the map description added as context.

  3. Carry on as usual — choose your focus, review the coverage map, then review and add the test cases. See The Coverage Map.


Journeys feed generation: a journey gives the generator the whole path — the steps in order and the persona walking them. It’s the best way to get end-to-end test cases: one case that crosses four screens instead of four cases that each test one.

  1. Select Journeys in the Navigate group of the action rail.

  2. Open the ⋮ menu on the journey’s card and choose Generate test cases…

  3. Continue in the generator as you would for a node.


A run started from Atlas uses Canon like any other, with the same Ground in Canon toggle.


Generated test cases go through the usual review, land in the folder you choose, and are ordinary test cases afterwards. To have them count towards the item’s coverage, save them in the test folder that item is linked to (or a sub-folder of it), then select Recompute coverage.