Generate test cases from requirements and links

Generate test cases with AI in Test Management from requirement documents, prompts, or links to Jira, Azure DevOps, Linear, Confluence, and Figma.

BrowserStack AI generates comprehensive test cases directly from your requirement documents, designs, or issue trackers. The agent analyzes your inputs to create detailed steps, preconditions, and expected results quickly.

How to generate test cases

You can generate test cases in a single flow, regardless of whether your requirements are in a PDF, a Jira ticket, or a Figma design.

Choose the destination folder

Accepted test cases are saved to the folder shown at the top of the review screen. To change where they are saved, do the following:

1. Click the edit (pencil) icon next to the folder name on the review screen.

   

2. In the Select destination folder dialog, select an existing folder, or click Create Folder to create a new one.

3. Click Set Location.

   

To create a separate sub-folder for each scenario under the destination folder, select Save scenarios as sub-folders from the dropdown next to the folder name.

Filter the generated test cases

To review a subset of the generated test cases, click the filter icon on the review screen. You can filter by testing type, tags, or priority, and choose whether the results match all filters or any filter.

You have now generated and added AI-powered test cases to your test case repository, linked directly to your requirement. Each test case that you add uses the default State value configured for test cases in your project. This applies to all templates, including BDD and custom templates.

Examples of generated test cases

The following image shows an example of an AI-generated test case along with its details.

The following image shows an example of an AI-generated BDD test case along with its details.

Persistent background generation

After you start the generation, the AI test case generation runs in the background. You can close the generation window or navigate to other parts of the application without interrupting the process. The job is tied to your current browser session. It stops only if you refresh the page or log out. To check the progress, reopen the Generate Test Cases with AI panel.

External context sources

External context sources bring requirement and design data from third-party platforms into your test case generation. By connecting these sources, you can link Figma frames, Confluence specifications, and documents from SharePoint or OneDrive to your test cases. You can also link issues from Jira, Azure DevOps, or Linear. Each generated test case then traces back to its source, from design to execution. The following table lists the supported integration types and their primary use cases.

Document limitations

BrowserStack AI performs best with PRDs that are well-structured, detailed, and unambiguous. For optimal performance, adhere to the following PRD constraints:

If your document exceeds the size or page limit, the upload will fail. To proceed, you must split the document into smaller parts or compress its content before uploading.

Scope your generation to part of a document

You do not have to generate test cases for an entire requirement document. To cover only a part of it, mention the specific section or line in your prompt. For example, prompt Generate test cases only for the Payments section of the attached document. The agent then generates test cases for that part of the requirement only.

Tool-level limitations

To ensure platform stability and consistent performance across all workspaces, the system enforces specific technical boundaries. These constraints define the maximum allowable values for data volume, and concurrent processes. Review the following table to ensure your automation suites and integrations operate within these supported thresholds.

Traceability and source retention

• On the review screen, the link icon next to each generated test case opens its source references. You can see exactly which part of the document, Jira issue, or Confluence page was used to create that test case, with the primary source, its section, and the excerpt used.
• Each generated test case stores the original Jira issue or Confluence URL as a reference attribute.
• The link remains clickable, so you can jump back to the exact source page at any time for clarification or updates.
• If the Jira issue or Confluence page changes later, you can re-run generation with the same link to produce updated or supplemental coverage.

Best practices

• Provide the most specific page (avoid large index/parent pages with broad, unrelated content).
• Pair a concise guiding prompt with the link if the page is lengthy or covers multiple modules.
• After major edits to the Confluence page, regenerate or manually review impacted test cases to keep them aligned.

Limitations

• Private or restricted pages require that the authenticated Atlassian user has at least view permission.
• Very large pages may be truncated to a maximum processed size. The system automatically prioritizes structured sections first.
• Embedded diagrams or images are not semantically parsed yet. Add clarifying prompt text if critical flows are only in images.

Next steps

• Refine the generated test cases with iterative prompts.
• For answers to common questions, refer to the test case generator FAQs.