AI Generated Test Cases

Testmo’s AI Test Case Generation is designed to fit into a complete, end-to-end testing workflow — from requirements, to test design, to execution, coverage, and reporting — all in one place.

Rather than treating AI as a standalone feature, Testmo integrates test case generation directly into your repository and requirements model, helping teams move from intent to coverage without switching tools or breaking traceability.

⚠️ Note: AI Test Case Generation is currently available as a BETA feature and continues to evolve based on customer feedback.

End-to-End Overview

At a high level, the workflow looks like this:

1. Start test case generation from the Testmo repository
2. Provide clear requirements as free-text input 
3. Review and guide AI-suggested test cases 
4. Generate final test cases and refine them as needed 
5. Save generated test cases to the repository with: 
   1. An AI tag 
   2. A link back to the originating requirement 
6. Track coverage, traceability, and execution progress across Testmo dashboards and reports 
7. Re-use requirements for future test case generation workflows

Each step is designed to preserve structure, ownership, and traceability - with AI assisting, not replacing, test design decisions.

Getting started: initiating test case generation

AI Test Case Generation is initiated directly from the Testmo repository.

To begin:

1. Navigate to your project’s Repository
2. Click Generate (BETA) in the repository header bar

Supplying requirements

Test case generation starts with supplying some requirements for the AI to work with. Additionally, you may select:

• The destination template to be used for your finished test cases. AI test case generation supports the following canonical test case types: Case (text), Case (steps) Case (BDD). You can use built-in templates or custom templates, as long as they conform to one of these canonical structures.
• The number of test cases to be created. You can request between 10 and 60 test cases per generation. Because AI generation is probabilistic, the exact number of returned test cases may vary slightly.
• The destination folder for your test cases. When starting AI test case generation, you’ll be asked to explicitly select an existing folder or create a new one. Folder context is not inferred automatically, ensuring generated test cases are organised intentionally from the start.

Requirements are provided as free-text input using Testmo's built-in rich-text editor. Attachments such as documents or images are not supported at this time.

Best practices for writing requirements

Clear, well-structured requirements produce significantly better test case suggestions. When writing requirements:

1. Give each requirement a clear, descriptive title. Requirement titles are used when linking test cases, browsing previous requirements, and viewing requirement references in runs and reports.
2. Describe system behaviour, not just features 
3. Include: 
   1. Preconditions 
   2. User actions 
   3. Expected outcomes 
4. Break complex functionality into logical sections 
5. Avoid vague language such as “handles correctly” or “works as expected” 
6. Focus on what should be tested, not how it is implemented

If the initial AI suggestions don’t meet expectations, you can return to this step, refine the requirements, and generate suggestions again.

Selecting & editing suggested test cases

Once requirements are submitted, Testmo generates an initial set of suggested test cases. 

This is a review and guidance step, not final generation. It's your opportunity to refine the suggested test cases, before they get finalised.

Guiding the AI

At this stage, you can:

• Edit test case titles and content 
• Deselect suggestions you don’t want to proceed with 
• Refine wording to better reflect edge cases or business rules

You cannot create entirely new test cases at this step.

If the overall direction isn’t right, you can go back, adjust the requirements (if necessary), and request a new set of suggestions.

This step ensures that human judgement shapes the final outcome before test cases are generated.

Refining & adding test cases to the repository

After reviewing and confirming the suggested test cases, you proceed to generate the final test cases.

This is a one-pass generation step:

• The AI uses your reviewed suggestions as guidance 
• Final test cases are created 
• You can further edit test case content before saving

Once saved, the test cases become standard Testmo test cases — fully editable, executable, and reportable.

Repository integration, traceability, and coverage

When test cases are saved:

• They are added directly to the Testmo repository 
• They are tagged with AI 
• They are linked to the originating requirement

This linkage enables:

• Requirement-to-test traceability
• Coverage tracking across your test suite
• Reporting via Testmo dashboards
• Visibility in the Coverage & Traceability report, where requirements and issues are tracked separately with independent controls
• Requirement references can be viewed, added, and removed from the test case edit and bulk edit dialogs
• Requirements can be linked to runs directly from the create/edit run dialog

Because requirement links are automatically added to generated test cases, they can be used to filter cases in the repository, select cases when creating runs, and generate requirements & traceability reports.

Rather than existing as isolated AI output, generated test cases become part of a connected testing model that spans design, execution, and reporting.

Re-using requirements for future generation

Requirements used for test case generation are stored for future use.

During a new test case generation workflow, you can:

• Browse previously used requirements 
• Select and re-use an existing requirement specification

Requirements can currently only be accessed and selected within the Test Case Generation workflow. They cannot be edited outside of it. 

When selecting previous requirements, Testmo shows how many test cases are already associated with each requirement, and allows you to navigate directly to them. Existing requirements can be reused as-is or modified to generate a new set of test cases.

This allows teams to:

• Standardise requirement inputs 
• Re-generate test cases as systems evolve 
• Maintain consistency across test design efforts over time

How this fits into a holistic testing strategy

AI Test Case Generation in Testmo is not intended to be a standalone feature or a “more is more” collection of AI capabilities.

Instead, it is designed to support a coherent testing lifecycle:

• Requirements define intent 
• AI accelerates test design 
• Human review ensures quality and relevance 
• Test cases live in the repository 
• Execution, coverage, and traceability are tracked in one system

By keeping requirements, test cases, execution results, and reporting connected inside Testmo, teams can reduce tool fragmentation and maintain a clear line of sight from what needs to be tested to what has been tested.

⚠️ Note: After completing (or abandoning) a test case generation workflow, users are prompted to provide feedback. This helps improve generation quality and identify where the workflow may need refinement.

Permissions & administration

AI test case generation can be disabled entirely via the admin console. Access can also be controlled via role permissions. Users must have permission to create cases and folders in order to generate test cases using AI.

FAQ's: Templates and AI-Generated Test Cases

How does the AI use my test case template?

When you select a template in the AI test case generation wizard, the AI scans that template for specific supported fields and fills them in automatically. Only fields it recognises are populated - all other template fields are left empty.

Which fields does the AI fill in?

The AI recognises fields by their system name. The supported system names are:

The AI determines which type of test case to generate based on the combination of supported fields present in your template:

Only templates that match one of these exact combinations will be available to select in the wizard. If a template contains a mix of fields that spans more than one type - for example, description +
 expected + steps - it will not appear as an option, because the AI cannot determine whether to generate a text-based or steps-based test case.

What about custom fields? Can the AI fill those in?

Not currently. The AI only populates the supported system fields listed above. Custom fields - including custom text fields - are not populated regardless of their name or configuration.

My template has required custom fields. The AI leaves them blank and I can't save the test case. What should I do?

As a workaround, add default text to any required custom fields directly in the template. The AI will preserve default values already set on a template, so this prevents the "required field is empty" error when saving generated test cases.

To do this: edit your template, find the required custom field, and enter a placeholder default value (e.g. N/A, TBD, or a relevant prompt for the tester to complete).

My template has a field with a supported system name (like steps) but the AI wizard shows an error when creating test cases. What's happening?

The AI wizard currently matches fields by system name only, without checking that the field type matches. For example, if your template has a field named steps that is configured as a plain Text type rather than a Steps type, the wizard will still try to use it - and will fail at creation time with an error.

Workaround: Check that any field using a supported system name (description, expected, steps, bdddescription) is configured with the correct field type. If the field name and type don't match expectations, rename the field or change its type to align with the table above.

The AI isn't filling in any fields at all. What should I check?

• Confirm the template you selected contains at least one field with a supported system name (`description`, `expected`, `steps`, or `bdddescription`).
• Confirm that field is set to the correct type (see table above).
• If you're not sure of a field's system name, check the template settings - the system name is shown when editing or creating a field.
• Alternatively, if you're working with the API, you can retrieve field system names programmatically using the Fields endpoint (GET /api/v1/projects/{project_id}/fields), which returns the system_name for every field in your project.

What’s next

AI Test Case Generation is currently in BETA, and we’re actively refining the experience based on real-world usage and feedback.