Note: This feature will be released in April 2024.

Testing engineers and QA engineers manually write and validate test cases, which is time-consuming, impacts accuracy, and could require multiple to-and-fro reviews for updates.

Freshservice now leverages AI to automatically formulate test cases from task titles & descriptions. This is inclusive as part of our Freddy Copilot offering.

Every user story on any task detail page has a section displaying Test Case Results. In the same section, users can now generate test cases to speed up end-to-end QA and testing of test cases. This is available for both custom and default task types.

Here’s how you can generate test cases automatically:

  1. In the Test Cases section, click on ‘Generate Test Case’. Here, the test case title and description are provided as input to the Generative AI-powered Freddy Copilot. Users can choose to input field values like ‘severity’ and ‘type’ as well.

  1. A CSV file with generated test cases will be made available under ‘Test Case Version Summary’. Various versions are displayed with timestamps so that the users can track the latest ones.

  1. Here, the user can choose to preview the test cases in-app as shown below.

3) The user can download the file with newly generated test cases. Test engineers can review and edit test cases as desired.

4)  After making changes, the file can be uploaded into the test cases module after mapping fields from CSV to the test case form.

Important Note: Test cases will be lost whenever a browser session is refreshed. Hence please download the CSVs that are generated, so there is no loss.

Here are a few guidelines for Generative AI usage:

Do’s - While using generative AI

  • Always review all output from Generative AI and refine it as needed.

  • Utilize Generative AI for various content-specific tasks like test case generation, and test data generation.

  • Provide additional context in the description to get more coverage.

  • The prompt passed to Generative AI is displayed and can be modified for further revisions to the description.

Don’ts - While using generative AI

  • Don’t include any development code in the prompt to perform white box testing to find the flaws in the development code.

  • If the task description has any sensitive information like the architecture of product / internal server details then those details should be removed as part of prompt preparation.

  • Review all test cases derived from Generative AI before further use. 

Learn how to create and manage test cases, test runs, and test plans in Freshservice here.