Running Autonomous Exploration

Start an Autonomous Exploration run in Test Manager against a requirement, with a time-boxed session, an AI model, and optional live streaming.

This task starts Autopilot exploring an application against a requirement, without a pre-defined test script.

Prerequisites

• A cloud robot, or an attached robot. Serverless robots are not supported.
• The Execute Automated Tests permission in Test Manager, required to trigger an exploration.

Steps

1. Log in to Test Manager.
2. Open a requirement.
3. Select + Autonomous Exploration.
4. Attach requirement documents or additional context, if needed.
5. Enter the target application URL.
6. Optional: Provide a custom prompt from the Prompt Library, used as User Instructions.
7. Set a time-box.
8. Select an artificial intelligence (AI) model.
9. Optional: Enable live streaming.
10. Select Start to begin the exploration.
11. Once the status changes to Completed, select Show report to view the full results.

Result

Autopilot begins exploring the target application. Findings populate as scenarios complete, and the run can be paused, guided, skipped, resumed, or stopped, as described in Human-in-the-Loop for Autonomous Exploration. After the live phase ends, the report becomes available for review.

Recommended prompt structure

A well-structured prompt for Autonomous Exploration typically includes:

• Mission
• Scope
• Focus areas
• Test ideas
• Risks and questions
• Timebox

Consider referencing classic exploratory-testing tour styles, such as the Landmark Tour, Saboteur Tour, Couch Potato Tour, or Money Tour.