Keynote: 10x Software Testing

Kristil Kruustuk at Agile Testing Days 2023 • 2 min

Notes taken on Nov 14, 2023

As software testing becomes increasingly complex, companies are turning to artificial intelligence (AI) to optimize their testing processes.

In this keynote, I will share my experiences as a software testing entrepreneur who has experimented with and integrated AI into our workflows to speed up testing processes and maintain high-quality standards for our global clients.

Attendees will learn how AI can transform test case management, issue resolution, and compliance, and get real-world examples of how AI can drive testing efficiency and accuracy. Join me to discover the potential of AI to revolutionize the software testing industry and learn how to effectively leverage AI in your own testing workflows.

Is about how we as people take little steps to bring higher quality. It does not happen over one night. It is gradually taking little steps every day.

10x software tester is somebody who went through different stages as a software tester.

Biggest myth: Anyone can do software testing 🙄

Back in the days, testing was about meeting the requirements.

Challenges: speed and coverage

A lot was manual testing

Then testing at the speed of light -> test automation

But … with automation there are also things that does not work

-> need for test automation maintenance because products change all the time

Outlook Mobile: founder was against automation. The product was changing so rapidly that automation did not make sense. Every single day they went over the feedback on the app stores.

Testing early, often, and everywhere -> CI/CD

Now, tailored tests for diverse demands

-> security testing, accessibility testing, UX, …

=> software testing is becoming more and more expensive

But becomes more and more important. Because everybody wants quality. When something goes wrong it can damage the reputation. Used to be 12% of the IT budget, now 18-19%

Increasing efficiency -> AI

Understanding our challenges

  • efficiency
  • quality
  • scalability

Generative AI sparked curiosity at Testlio

-> ChatGPT unlocked progression

=> try to improve your processes

ChatGPT can generate test ideas, it gives a lot of great answers, not perfect, but a lot better

  • test creation
  • test refactoring
  • issue refactoring => 15% to 30% faster

it does not need to be perfect, it is these little steps that make the difference

AI helps with issue submission

=> 46% improvement in fixing errors in issue submissions

ChatGPT helps with the set up of test workspaces: how to set up the test framework

The rollercoaster of AI in testing

  • AI has an unpredictable nature: the algorithm changes resulting in different
  • the security tightrope: we don’t want to share confidential data with ChatGPT
  • the waiting game of response times: sometimes within 1 sec, sometimes within 10 sec
  • the formatting puzzle
  • prompt engineering

Key takeaways

AI has given us a lot of room for improvement

When we started playing with AI we realised it wouldn’t happen over night

Adaptability is key in our ever changing testing industry

Human expertise is essential for AI’s potential

Testers won’t be replaced by AI but they will be replaced by testers working with AI