Skip to content
Home » MagicPod releases two new functions using generated AI: “Test case summary function” and “Test change summary function”

MagicPod releases two new functions using generated AI: “Test case summary function” and “Test change summary function”

MagicPod releases two new functions using generated AI: “Test case summary function” and “Test change summary function”

*View in browser* *MagicPod Co., Ltd.*
Press release: June 27, 2024
MagicPod releases two new functions using generated AI: “Test case summary function” and “Test change summary function”
*Tips for improving the accuracy of summaries and prompts used internally are also revealed! *
MagicPod Co., Ltd. (Headquarters: Chuo-ku, Tokyo, Representative Director: Nozomu Ito) is an AI test automation platform.
The new features of MagicPod are the test case summarization function and test change summary function using generation AI.
We would like to inform you that we have released “Noh”.
New feature details
Test case summary function
When you create a test using MagicPod, it automatically generates a summary of what the test is about.
Until now, in order for another person to understand the contents of a test created by one person, it was necessary to read the test contents.
Questions that require time to understand or difficult to grasp the overall picture because it requires detailed explanation or asking the creator.
There was a problem. By automatically creating a test summary, even people other than the person who created the test can easily and quickly understand the content of the test just by reading the summary, making it easier to share information. Furthermore, when the content of the test is updated, the summary is automatically updated, so there is no need for humans to update the summary, and there is no risk that the content of the summary will become outdated compared to the test.
Automatically generate Japanese summaries for test cases where you don’t know what they are doing at first glance
Convert test scripts to human-readable text with MagicPod Public API and summarize it with generative AI
Test change summary function
When modifying and saving a test case, a draft “change message” summarizing the changes can be created with the push of a button. It can now be generated. The generated change message proposal can also be modified by the user if necessary.
Until now, to make it easier to trace the change history of a test, humans had to come up with easy-to-understand change messages. However, by using generation AI, the time required for this work can be reduced. vinegar. Additionally, by making it easier to create change messages, more users can write meaningful change messages.
history is easier to track.
Use AI to generate and save a summary of changes that are difficult to understand at first glance
Get the before and after test scripts as human-readable text using the MagicPod Public API and summarize it using the generative AI. About learning data
The training settings of the generative AI are disabled, and information such as customer test scripts is used to train the generative AI.
You will not be disappointed.
How to Use
Users who subscribe to MagicPod’s paid plans can use it without any additional charges, regardless of the plan details.
I’ll take it. However, in order to enable the function, advance application is required before use. Users who wish to use this service should log in to MagicPod and send the following information using the “Contact Us” form at the bottom right of the screen within the service.
– Title: Application for use of generation AI function
– Body: MagicPod organization name to be used
To confirm your identity, you will need to apply using the form within the service after logging in.
Please note that once the generation AI function is enabled for an organization, it cannot be disabled again.
Please note that it may take 1-3 business days for activation. How to write MagicPod tests to improve summary accuracy
By making the test script easy for humans to understand, it also becomes easy for the generation AI to handle.
This increases the accuracy of the summary. Specifically, the following three are effective.

Divide the test script into several sections separated by blank lines, and include a summary comment on the first line of each section that describes what that section is doing.

If the screen element name used in the test step is difficult to understand, such as “Area 1”, change it to an easy-to-read element name such as “Register Button”.
– Use “Japanese” or “English” in the same test script
Example script that improves summary accuracy
Publish prompt
We will publish the prompts used internally for the above new features. This prompt should be run with ChatGPT3.5
It is assumed that.
Prompts used in “Test case summary function”
Please briefly (up to 300 characters) summarize the user flow of the browser test below in Japanese. :
[Test script in text format]
Prompts used in the Test Change Summary feature
Below are the contents of the browser test script before and after the change.

Change before: