Generative AI in Software Testing: How testRigor Generates Tests For You

Summary

This tutorial demonstrates the ease of using generative AI with testRigor to create end-to-end tests. By using BestBuy as an example, the video shows how to set up a test suite, generate test cases, and modify them for accuracy. The AI autonomously navigates the website, identifies steps for a functional test, and allows for quick edits to ensure the test meets specific requirements. The result is an efficient and customized test case in just minutes, showcasing the power of AI in simplifying the testing process.

Takeaways

• 😀 The video demonstrates the ease of using generative AI to create end-to-end tests with testRigor.
• 🛍️ The demonstration uses the BestBuy website as an example to show the test creation process.
• 📝 The user is instructed to copy and paste the website URL into the test feed and name the test suite.
• 🔄 There is an option to add custom test cases manually or let the AI generate them based on the website's content.
• 🔑 The AI uses prompts and the website to determine step-by-step actions needed to build a functional test.
• 🔍 The system automatically loads the specified website and begins the test generation process.
• 📸 The video shows a screenshot of the AI entering 'Kindle' into the search bar and adding it to the cart.
• 🛒 The AI navigates through the checkout process, including handling a login prompt and entering an address.
• ✏️ The test case can be easily modified using Quick Edit or the Edit test case button for corrections.
• 📋 The AI-generated test includes validation steps, which can be adjusted based on the user's needs.
• 🔄 After modifications, the AI retests the updated instructions to ensure the test performs as required.
• 🎯 The video concludes with a functioning negative end-to-end checkout flow test created in under five minutes.

Q & A

What is the main topic of the video script?

The main topic of the video script is demonstrating how to use generative AI to create end-to-end tests with testRigor, using BestBuy as an example.

What is the first step in creating a test suite in the video?

The first step is to copy the URL, paste it into the test feed, and name the test suite, in this case, 'BestBuy'.

How does the video script suggest adding a custom test case?

The script suggests clicking on the 'Add custom test case' button and then writing the test cases using description and steps, or using the system to generate an example.

What does the system do when the second button is clicked for generating a test?

When the second button is clicked, the system starts generating an actual functional end-to-end test using the website URL specified earlier.

What is the purpose of the prompt used by the system in the video?

The prompt is used by the system to figure out step by step what needs to be done to build a functioning end-to-end test on the specific website.

How does the AI system interact with the website during the test generation?

The AI system loads the specified website, enters information into search bars, clicks buttons, and navigates through pages to simulate user actions for test generation.

What issue did the AI encounter when entering an address during the test?

The AI encountered an issue where it entered the full address into the address bar, which is not in the correct format because city and zip code are different fields.

How can the generated test case be modified if there are errors?

The test case can be modified using the 'Quick Edit' or 'Edit test case' button to correct any errors or adjust the steps as needed.

What validation method is suggested in the script for the test case?

The script suggests validating the test by checking if the page still contains the 'Apply' button after making changes to the test case.

How long did it take to create a functioning negative end-to-end checkout flow test according to the video?

It took about five minutes to create a functioning negative end-to-end checkout flow test, with most of the test case generated by the generative AI.

What is the final outcome of the test case generation process shown in the video?

The final outcome is a functioning test case that is working and performing exactly as needed, with the majority of the test case generated by the AI and modified by the user for specific requirements.

Outlines

00:00

🛒 Generating E2E Test Cases with AI

The video demonstrates the process of using generative AI to create end-to-end test cases for a website, specifically BestBuy. The presenter guides viewers through copying the website URL, naming the test suite, and utilizing the 'Add custom test case' feature. Two options are presented: writing test cases manually or allowing the system to generate them automatically. The focus is on the latter, which involves the AI loading the website, using prompts to understand the steps necessary for a functional test, and executing those steps. The AI's process is shown, including entering search terms, adding items to a cart, and navigating through checkout, even handling a login prompt and address entry. The presenter also shows how to modify the test case using 'Quick Edit' to correct any issues, such as entering an address in the wrong format.

05:05

🔍 Validating and Updating Test Cases

This paragraph continues the tutorial by discussing how to validate errors and specific conditions within the test case. The presenter explains how to highlight necessary fields like the State or to ensure that the 'Apply' button remains functional. The video shows the updated test case execution, which follows the corrected instructions provided by the user. The AI's generative capabilities are emphasized, as the majority of the test case is automatically generated, with the user making necessary modifications to fit specific needs. The video concludes by showcasing the successful execution of the updated test case, which validates the 'Apply' button's functionality, demonstrating the ease of building functional end-to-end tests using AI.

Keywords

💡generative AI

Generative AI refers to artificial intelligence systems that are capable of creating new content, such as text, images, or data, that resembles the patterns found in existing data. In the context of the video, generative AI is used to automatically generate test cases for website functionality, showcasing its ability to understand and replicate the process of creating end-to-end tests.

💡testRigor

TestRigor is a testing platform mentioned in the script, which is used for creating and managing automated tests. The video demonstrates how to utilize TestRigor in conjunction with generative AI to generate tests for a website, emphasizing the ease and efficiency of this process.

💡test suite

A test suite is a collection of test cases that are designed to be executed together to determine if a particular software meets its design and behaves as expected. In the video, the test suite is named 'BestBuy', and it serves as the framework within which the AI generates and modifies test cases.

💡custom test case

A custom test case is a test scenario that is specifically designed to meet the unique requirements of a particular software testing need. In the script, the option to add a custom test case allows the user to manually input descriptions and steps for the AI to generate a test based on these specifications.

💡end-to-end test

An end-to-end test is a type of test that covers all the stages of a process or system to ensure that it functions correctly from start to finish. The video demonstrates how generative AI can create an end-to-end test for a shopping cart and checkout process on a website.

💡prompt

In the context of the video, a prompt is a set of instructions or a cue that the AI uses to understand the task at hand. The AI uses the website URL and a prompt to determine the steps needed to build a functional test.

💡screenshot

A screenshot is an image captured by a software tool to record the contents displayed on a computer screen at a particular moment. In the script, the user clicks on a screenshot to observe the actions performed by the AI during the test generation process.

💡Quick Edit

Quick Edit is a feature that allows users to make immediate changes to a test case without going through a separate editing interface. In the video, the user utilizes Quick Edit to modify the test case by removing unnecessary validation steps and correcting the address format.

💡validation

Validation in software testing is the process of checking the correctness, completeness, and consistency of a software application. In the video, the AI initially includes a validation step to check if the page contains an address, which the user later modifies.

💡retest

Retesting is the process of executing a test again after changes have been made to the software or the test environment. In the script, the user updates the test case and then retests it to ensure that the modifications have been correctly implemented.

💡checkout flow

Checkout flow refers to the sequence of steps a user goes through to complete a purchase online. The video demonstrates the creation of a test case for a negative checkout flow, which would simulate scenarios where the checkout process might fail or encounter issues.

Highlights

Introduction to using generative AI for creating end-to-end tests with testRigor.

Demonstration of creating a test suite by copying a URL and naming it BestBuy.

Explanation of adding a custom test case manually or generating one automatically.

Showcasing the automatic generation of an example test case.

Emphasizing the generation of actual functional end-to-end test using the website.

Description of the system loading the specified website to build a functioning test.

Observation of the AI entering 'Kindle' into the search bar and adding it to the cart.

Illustration of the AI navigating to the cart and proceeding to checkout.

AI's decision to continue as a guest during the login prompt.

Misstep in the AI's attempt to validate the page contains an address.

Correction of the AI's incorrect address entry by modifying the test case.

Introduction of Quick Edit feature for easily modifying test cases.

Explanation of validating errors and specific conditions in the test.

Demonstration of updating and retesting the modified test case.

Showcasing the updated test executing the correct instructions.

Achieving a functioning negative end-to-end checkout flow test in minutes.

Majority of the test case generated by generative AI with user modifications.

Final demonstration of the test case working as needed and validating the 'Apply' button.

Conclusion emphasizing the ease of building functional end-to-end tests with AI.

Transcripts

Hello everyone! Today I will show you how easy it  is to use generative AI to generate end-to-end  

tests using testRigor. For that we will be using  BestBuy. Let's just copy the URL, paste it into  

the test feed and name test suite BestBuy,  then we can proceed and create the test suite.  

There is a button here says Add custom test  case. You click on that: here you can write  

your test cases yourself using description  and then steps that need to be executed. Or  

you can specify the description for example,  find a Kindle and add it to shopping cart.  

If you click on the first button, it  will just generate you an example.  

However, if you click on the second button,  

this is where it will actually start generating  your actual functional end to end test,  

and this is automatically generated test in  there. So we are interested in the second one.  

So what the system does is basically it loads  the website, the URL which we have specified  

when we created the test suite. Then  it will use a prompt and your website  

to figure outstep by step what needs to  be done to actually build a functioning  

end to end test on your specific system.  And this is what it is doing right now.  

So as you can see it entered, let's go there  and see what it is trying to do here. If I  

click on the screenshot you can see it entered  Kindle into search bar, clicked on this Kindle  

and Add to cart, and it got  to this interesting pop up.  

But it figured out that you  should click Go to cart there,  

and it got to whatever seems to be Cart. Then  

it clicked Check out and there's a login  prompt.It says continue as guest. All right  

here it decided for some reason that it needs  to validate that the page contains address.  

After that it actually entered some kind of  an address there. As you can see it tried to  

enter the full address into address bar which  is clearly not to the format because the city  

and the zip code are different fields. But that's  totally fine because we can easily modify this  

test case. For that you can use either Quick Edit  or Edit test case button. Let's click Quick Edit,  

and here first of all, we do not need to do  this validation in here so we'll remove that  

we might need it later. Second  of all, it's a funny city name.  

We can add this city into City and enter the  zip code into Zip and remove them from the  

original long address. Now we are entering  correct format, address city and zip,  

click Apply which is next button in  there. So now the question is, okay,  

we're probably not going to go through, what are  we going to validate? Well, there are several  

way show we can validate errors we can validate  specific errors. Let's say it highlights the State  

and says we need the State, or we can  just validate it and go through anywhere,  

and say the page still contains Apply  button and this is what we'll do,  

so just click update and retest. As you  can see here, this is our updated test.  

And now what it will do, it will  execute our updated instructions  

to run the test doing exactly what we need there  instead of these slightly incorrect ones. And this  

way we got ourselves a functioning negative end  to end checkout flow test within five minutes.  

Once again, the majority of the test case  had been generated by the generative AI.  

And then we modified it to  allow it to do whatever we need.  

And we got ourselves a test case which is  working and doing exactly what we need.  

And this is how easy it is to build functional  end to end test case using generative AI.  

Let's give it a second and voila!  As you can see here, it found Apply  

which is still enabled button on  the page. Thank you very much!