Autonomous Testing - How this can help your testing work?

We asked ChatGPT to write a blog about introducing autonomous testing to manual testers:

Here is the results:

In today’s fast-paced software development industry, there is a growing demand for faster and more efficient testing processes. One way to meet this demand is through autonomous testing, which involves the use of artificial intelligence and machine learning to automate various testing tasks. However, many manual testers may feel intimidated by the concept of autonomous testing, fearing that it may eventually replace their jobs. In this blog, we will discuss how to introduce autonomous testing to manual testers in a way that will help them understand the benefits of this technology and how it can help them in their roles.

  1. Explain the concept of autonomous testing

Before introducing autonomous testing to manual testers, it’s important to explain the concept and benefits of this technology. Autonomous testing uses artificial intelligence and machine learning algorithms to automate repetitive and time-consuming testing tasks. This can help testers save time and reduce the risk of human error, allowing them to focus on more complex and challenging testing scenarios.

  1. Address concerns and answer questions

Many manual testers may feel intimidated by the idea of autonomous testing, fearing that it may replace their jobs. It’s important to address these concerns and answer any questions they may have. Explain that autonomous testing is not meant to replace human testers, but rather to augment their work and allow them to focus on more critical tasks. Reassure them that their roles are still valuable and necessary in the testing process.

  1. Provide training and support

To ensure that manual testers feel comfortable with autonomous testing, it’s important to provide them with proper training and support. This can include workshops, webinars, and other training sessions that focus on the fundamentals of autonomous testing and how it can be integrated into their existing workflows. Provide them with the necessary tools and resources they need to use autonomous testing effectively.

  1. Start small and gradually scale up

When introducing autonomous testing, it’s important to start small and gradually scale up. Begin by automating small and repetitive tasks, such as regression testing, and gradually move on to more complex scenarios. This will allow manual testers to gradually become familiar with autonomous testing and gain confidence in their ability to use it effectively.

  1. Encourage collaboration

Autonomous testing should not be seen as a replacement for manual testing, but rather as a complement to it. Encourage collaboration between manual testers and autonomous testing tools, as this will help to maximize the benefits of both approaches. By working together, manual testers can use their expertise to identify critical issues and provide feedback to improve the accuracy and effectiveness of autonomous testing.

In conclusion, introducing autonomous testing to manual testers can help to improve the overall efficiency and effectiveness of the testing process. By explaining the concept and benefits of this technology, addressing concerns, providing training and support, starting small and gradually scaling up, and encouraging collaboration, manual testers can gain confidence in their ability to use autonomous testing effectively and see the benefits of this technology in their daily work.

It provides general knowledge for those who are interested in this topic, and also, I’m curious to know more about the power of this technology in testing, a self-learn post :smiley:

Read more at Katalon Blog: Autonomous Testing: A Complete Guide | Katalon

This Katalon Community web site would deserve AI capability. AI will be a great help if it coud do triage for the questions from users.

In this forum we have a lot of repetitive questions from new users. Those volunteer guys repeat reacting to them with reply that say:

  • Lack of concrete information. Have a look at the guideline
  • Please share the URL of your target web page if it is public. If it is private, save the HTML source into file and share it here.
  • You saw a message "Unable to locate element”. First, read the HTML source carefully. Second, check the locator of your Test Object is correct against the HTML source. Third, …
  • Do you have <iframe> element in your taget HTML? Please check it.
1 Like

Totally agree @kazurayam, we are super thankful to your contribution to all testers coming here jotting down their questions.

We are trying many ways to minimize the gateway tasks, improve search functionality and re-enforce guidelines on new users but sure AI could really help categorization and prioritization in this community forum.

Even @mike.verinder tried something today by asking ChatGPT to solve one of forum questions:

Q: I use katalon recorder 5.9.0 and test web site. While testing, I can see the erorrs on the site, but they are not shown in katalon recorder, all tests are passed. What should I do to chach errors in katalon recorder? Thanks!


If you’re not seeing any errors in Katalon Recorder, even though you’re seeing errors on the website, it’s possible that the errors are occurring after the tests have completed. Here are a few things you can try:

  1. Increase the timeout: Sometimes, errors can occur because elements take longer to load than expected. Try increasing the timeout for commands that interact with elements on the page.
  2. Use the verifyElementPresent command: The verifyElementPresent command checks if an element is present on the page. If it’s not present, the test fails. This can be a useful way to catch errors that occur when elements are missing.
  3. Check the console for errors: Even if Katalon Recorder doesn’t show any errors, the browser’s console might have useful information. Open the console (usually by pressing F12) and look for any error messages there.
  4. Use try-catch blocks: If you suspect that errors are occurring within a specific section of your test, you can wrap that section in a try-catch block. This will catch any errors that occur within the block and allow you to handle them appropriately.
  5. Debug your test: If all else fails, try debugging your test. Set breakpoints at key points in your test and step through it line by line to see where the errors are occurring.

Hopefully, one of these tips will help you catch the errors you’re experiencing. Good luck

But still, no one can help when vital information is not provided…


Why not you start with applying AI for Katalon Community now?

I would welcome if a new member joins this forum and replies to all of questions.

First, tell ChatGPT to RTFM carefully :slight_smile:

Sorry, what is RTFM?

I found RTFM in Wikipedia. Thank you. I hope the AI for Katalon Forum could reply RTFM to users smartly.

AI capability goes beyond testing. At some point in their evolution, software and the systems it runs on will become so complex and cross-integrated that they will only be able to be developed by AI-enabled agents – essentially, software developing itself. Hopefully, it will continue to keep its human beneficiaries in mind!


keep dreaming.
ar the moment of writting, any AI approach i know it is based on training.

trained on wrong resources will produce wrong results.

AI it is not self aware (and that is still a good thing) therefore have no actual understanding of ‘good’ vs ‘bad’

AI may be able to answer to a question like ‘who are you’ or ‘what are you’ but will be just the most convenient answer he learned during training.

is just using such terms based on the most expected answer criteria.

when AI will be self aware… well, ok. good luck with that, i hope at that time i will be passed away.

1 Like

@Russ_Thomas did some relevant experiments attempting to teach chatGPT to write propper code.

i will let him to conclude on this matter.

my challenge for AI fans is opposite.

ask the AI for whatever he will give a good solution (and here, the quality of the solution is relative, since you, as a person will judge)

once you get the ‘satisfactory’ solution, start to teach the AI the wrong way and insist in that.

Any sane human being will say at a moment ‘f*** you’

The AI will most probably keep chating and eventually accept your solution.

the challange is started, have fun!

Okay, here’s my conclusion:

When it looks like you’ve taught ChatGPT a lesson, you’re only fooling yourself. ChatGPT’s ability to learn from human users’ input is NOT enabled. I will not waste my time like that again :angry: :nerd_face:

Not sure I agree with all of that, but since it is written in such a loose/unfounded way, it’s hard to build an argument for or against any of it.

The author has not decided whether Autonomous Testing “is” (present tense) or “will be” (future). For any discerning reader (i.e. me), it leaves me wondering if I’m wasting time reading it.

Advice: Take a stand and say what you mean. Be forthright and diligent in your message. As it stands, this blog post is neither.


Ah … in the context of what I mentioned before, it may have sense why they did that.
So using it is actually a bet on if the AI was properly trained with correct sources.

I don’t see a usage for me in the near future for such tools.
I am too suspicious by the nature of my work.
I prefer human mistakes over software missbehaviour, it is easiest to understand what happened and how to fix it in case something went wrong.

I just used the Bing Chat bot (which apparently is now powered by ChatGPT-4) to generate a similar blog post to @vu.tran 's using the prompt: “Write me a blog to introduce autonomous testing to manual testers”, and below is the result:

Autonomous testing is the automated creation of tests, the maintenance of tests and analysis of the results with the computer acting and thinking like a human brain. This kind of autonomy equals faster and more efficient testing, which is why it is becoming increasingly popular in the software testing industry.

If you are a manual tester, you may be wondering how to get started with autonomous testing. The first step is to understand the benefits of autonomous testing. Autonomous testing can help you save time and money by automating repetitive tasks, reducing the risk of human error, and increasing the speed of testing.

The next step is to learn about the different types of autonomous testing. There are several types of autonomous testing, including:

  • Automated testing: This is the most common type of autonomous testing. It involves using tools to automate the creation, execution, and maintenance of tests.
  • Intelligent testing: This type of testing uses artificial intelligence (AI) and machine learning (ML) to create and execute tests. It can help you identify patterns and trends in your testing data, which can help you improve your testing process.
  • Autonomous testing: This is the most advanced type of autonomous testing. It involves using AI and ML to create, execute, and maintain tests. It can help you identify defects and issues that you may have missed with manual testing.

To get started with autonomous testing, you should start by learning about the different types of autonomous testing and the benefits of each type. You should also learn about the different tools and technologies that are available for autonomous testing.

In conclusion, autonomous testing is the future of software testing. It can help you save time and money, reduce the risk of human error, and increase the speed of testing. If you are a manual tester, you should start learning about autonomous testing and the different types of autonomous testing that are available.

I hope this blog has been helpful in introducing autonomous testing to manual testers. If you have any questions or comments, please feel free to leave them below.

Looking at the screenshot, at the bottom of Bing’s response there are 3 links. Now I’m not sure whether it only crawled through those articles and then reworded/paraphrased what was written in them, or that it crawled through numerous pages but only show those articles based on their SEO ranking.

Regardless, I think these “conversational search” experiences will create a new playground for marketers in terms of SEO and page rankings, let me know what you think.

1 Like