Understanding Agentic Learning and Run with Copilot

Testsigma provides two AI-powered features to improve the quality and reliability of test cases. While both features enhance the automation process, they serve distinct roles in the testing lifecycle. Agentic Learning assists in the design and discovery of the correct test flow, while Run with Copilot verifies that the test executes correctly within a real environment.

Run with Copilot

Run with Copilot validates a generated test case through real-time execution. This feature runs the test steps in a selected environment to ensure the automation is ready for deployment.

Capabilities

Use Run with Copilot to perform the following tasks:

• Execute generated test steps to identify failures.
• Identify issues with element detection.
• Validate that assertions and test data are correct.
• Debug steps before saving the test case to the library.

When to Use Run with Copilot

Select Run with Copilot to confirm that a test runs successfully from start to finish. Use this feature to verify the following conditions:

• The test runs without errors.
• The AI identifies all page elements correctly.
• The test data work as expected in the specific environment.

Agentic Learning

Agentic Learning identifies the correct test flow when information is missing or incorrect. It functions as an exploratory tester by learning the application's behavior and adjusting the test case to match the actual user journey.

Capabilities

Use Agentic Learning to perform the following tasks:

• Identify application navigation paths.
• Locate and add missing steps to the test flow.
• Correct steps that deviate from the expected workflow.
• Learn common behavior patterns within the application.

When to Use Agentic Learning

Select Agentic Learning in the following scenarios:

• The generated test flows are incomplete.
• Steps are missing or are in the incorrect order.
• The application navigation path is unclear.

Compare Run with Copilot and Agentic Learning

Use the following table to understand when to use Run with Copilot and Agentic Learning while generating and validating test cases: