Last updated Sep 20, 2026

Good practices to create Test Cases with AI

Struggling to identify how to begin writing test cases from a requirement is a common challenge. To address this, Requirements and Test Management (RTM) for Jira uses an Atlassian Rovo agent to generate initial Test Case drafts based on your requirement content.

Before you start

To use Test Case generation, make sure that:

Rovo is active on your Jira instance. The Create Test Cases (Rovo) button is only visible when Rovo is enabled. See Manage Rovo access for details.
Generating Test Cases with Rovo is turned on in RTM global settings under the AI tab. See AI settings for details.
Learn how RTM uses Rovo to generate Test Case drafts directly from your Requirements - see Generate Test Cases with Rovo.

How to generate Test Cases with Rovo

You can start generation from three places in RTM:

Requirement detail view - open a requirement, go to the Test Cases tab, and click Create Test Cases (Rovo).
Requirements list toolbar - select one or more requirements (up to 5) and click Create Test Cases (Rovo) in the toolbar.
Test Cases list toolbar - click Create Test Cases (Rovo) in the toolbar and select requirements using the picker (up to 5).

After you click the button, the Rovo agent opens in a panel. The agent reads the selected requirements and generates Test Case drafts. When the drafts are ready, a notification appears in the progress widget. Open the widget to review, accept, or reject each draft.

Note

If the Rovo panel does not open after you click the button, Rovo may not be active on this Jira instance even though the button is visible. Check the Rovo icon in the top-right corner of Jira or contact your administrator.

When should you use Rovo to generate Test Cases?

Use it as a first draft, not the final version. Many of our QA specialists use this feature as a starting point - a way to quickly generate a rough version of a Test Case that captures the idea. It helps when you’re short on time or unsure how to structure test steps. After generation, you can refine the output to suit your style, standards, and edge cases.

Start with clear, specific Requirements

AI uses natural language input to create your Test Case - if the input is unclear, the output will be too.

Tips for better input:

Use simple and clear language.
Avoid internal jargon (or explain it if needed).
Mention expected outcomes or key conditions.
Include relevant user roles or system states.

Example of effective prompt

As a registered user, I want to change my email address in the account settings so I can receive notifications on a new address.

Always review and improve the output

AI speeds up writing, but it doesn’t replace QA judgment. Think of the generated test case as a rough draft.

Tips for reviewing:

Add missing test data or preconditions.
Verify that test steps follow logical flow.
Manually include edge cases or negative paths.
Ensure expected results are clear and measurable.

Example before & after

AI-generated step: Click on ‘Reset Password’.

Improved: Click on ‘Reset Password’ and verify that a confirmation email is sent to the registered address.

Standardize titles and step formatting

Consistency makes tests easier to read, maintain, and filter - especially across large projects.

Tips:

Use consistent phrasing in actions (for example, ‘Click on’, ‘Enter’, ‘Verify that…’).
Apply team-wide naming patterns for test case titles.
Align with your test case template - adjust AI-generated structure if needed.

Example

Use: Verify password strength when registering.

Avoid: ‘Password check’ or ‘Test if password works’.

Treat AI-generated Test Cases as your first draft

Many QA teams use AI-generated test cases as the first step - to kick off the process and structure their thinking.

Why this helps:

You get a basic outline without starting from scratch.
It sparks ideas and highlights typical flows you can build on.
It speeds up early-stage work like test planning or exploratory prep.

Usage example from Deviniti

“Our team uses AI-generated tests as an initial draft. They help us quickly cover the ‘happy path’ and then we layer on negative scenarios and edge cases. It’s like getting a helpful skeleton that we flesh out based on our experience.”

Use AI as a pairing tool for Junior Testers

Junior QA specialists can use AI to learn about test case structure, common test flows, and phrasing. It’s a hands-on way to explore how requirements turn into test logic.

Common Mistakes to Avoid

Even the best tools can lead to problems if misused. Here’s what to watch out for:

Trusting AI output blindly - Every Test Case still needs human review - especially for logic, edge cases, and accuracy.
Generating too many irrelevant Test Cases - focus on quality over quantity. AI can create many cases quickly, but not all will be useful.
Skipping negative and edge case scenarios - AI-generated tests often focus on happy paths. Always add manual coverage for edge cases and failure conditions.
Giving vague prompts - input like ‘Login feature’ is too generic. Add specific user actions, expectations, or context.

Summary

Rovo-generated Test Cases in Requirements and Test Management (RTM) for Jira help you go from requirement to test structure faster - but the real power comes when you treat Rovo as your assistant, not a replacement. Use it to speed up ideation, draft better tests, and free up time for critical thinking and edge-case analysis.

Want to try it out?

Open any requirement in RTM, go to the Test Cases tab, and click Create Test Cases (Rovo).

Need help?

If you can’t find the answer you need in our documentation, raise a support request. Include as much information as possible to help our support team resolve your issue faster.