Ask questions about a test run, investigate failures, and report bugs.
Thunders AI is available inside the Test Run View. It uses the run's data (steps, results, screenshots, environment, and persona) to give contextual answers about what happened and what to do next.
You can also investigate a run from the Thunders AI tab: paste the test run ID and ask your question.
Open Thunders AI in a test run
In the Test Run View, click the Thunders AI button in the header bar, or use the keyboard shortcut Cmd+J on Mac or Ctrl+J on Windows.
The Thunders AI panel opens on the right side of the screen alongside the run details.
Suggested prompts
Thunders AI shows suggested prompt chips based on the run status. You can also type your own questions.
Examples of what you can ask:
"What is this test about?" Summarizes the test intent and the main flow being validated.
"Why did this run fail?" Analyzes the run results and explains the failure.
"Why did step 4 fail?" Focuses on a specific step, comparing expected versus actual results.
"Summarize this run." Provides a concise overview of the run outcome and key failures.
Investigate a failed step
When a step fails, its reasoning section shows a "Why did this step fail?" button. Clicking it:
Opens the Thunders AI panel (if not already open).
Attaches the failed step as context.
Starts the investigation scoped to that specific step.
Navigating between steps updates the context automatically. Selecting a non-failed step clears the chip and restores the suggested prompts.
Report bugs
When a tool is connected, you can ask Thunders AI to report a bug directly from a failed run or step.
Start from a failed step or run.
Choose the destination tool.
Fill in the guidance input.
Click Send.
The ticket includes the relevant failure context, such as the test case, run, failed step, expected result, and actual result.