Pega Gen AI Application Tests

AI
1

Hello Team,

For Pega CS constellation applications, can we showcase the steps required to create test cases using GenAI autopilot? consider how will interaction context and details will be passed to service case while generating test case for it.

2
  1. Access the Testing Workspace: Navigate to your Constellation-based application portal or App Studio, and select the automated testing or Autopilot landing page.
  2. Select the Target Component/Case Type: Identify the service case you want to test
  3. Trigger the Generator: Use the Pega GenAI Autopilot assistant to extract and analyze requirements, then select Generate Test Case to create the test scenarios.
  4. Define Context and Data: The assistant processes the requirements, personas, and associated steps.
3

I’d highly recommend this video: Autopilot for Testing: Accelerate Quality with AI‑Driven Test Automation | Pega by @AnthonyLeonardi.

It’s only 10 minute and well worth your time, if you are after a walk through.

We also have more on Search - Videos | Pega if you want to browse there.

4

Thanks for raising this — testing enablement is one of the best places to use GenAI as a bounded accelerator.

A predictable way to showcase it is:

  1. start from a clearly defined case type and happy-path scenario, ensure the interaction/service context you want reflected is captured as explicit test data (customer profile, channel, intent, key inputs),

  2. then use Autopilot to generate candidate tests and review them against your acceptance criteria before committing.

  3. Keep governance tight by treating generated tests as proposals: validate data setup, expected outcomes, and assertions, and version them like any other test asset so changes are auditable and repeatable.

If you can share which CS interaction elements must be passed (e.g., contact details, channel metadata, intent classification, policy outcomes), others can suggest concrete ways they model that context in test data for Constellation.

How have others structured their test data and review checkpoints to keep AI-generated tests reliable over time?