Testing and AI - now and in the future

Infotiv is now launching the concept of TestDev Powered by AI, where artificial intelligence is gradually being integrated into existing testing operations. We meet Maria Kindmark Alemyr, Head of Consulting at Infotiv, to explore how AI is changing the world of testing.

When Maria describes how AI is used in testing processes, it becomes clear that it is not about the future - but the present. The initiative is a concrete example of how Infotiv combines testing expertise with AI expertise to streamline customers' development processes.

What does TestDev Powered by AI mean and what problems do you help customers solve?

- The concept is really Infotiv in a nutshell. We have expertise in both testing and AI - now we're bringing them together. Customers want to get products to market faster, and our concept TestDev Powered by AI enables this. We have moved from manual to automated testing; now we are taking the next step in our journey and introducing AI to further streamline our customers' processes," Maria says.

- Through AI, we can increase the repeatability of our testing without using more resources. AI helps us with corner cases that humans easily miss and helps us find weaknesses in processes and strategies. AI also supports us in generating test scenarios and test cases, and helps us with requirements analysis and test coverage.

- But it takes sharp engineers with technical know-how and a critical eye to ensure the right requirements, test coverage and quality.

AI improves quality

Can AI on its own also improve quality?

- Yes, definitely. AI can analyze weaknesses from large data sets, identify where bugs have historically occurred, and can help us verify bug fixes by recreating scenarios and re-running tests.

- AI can also identify fixes in the code and update test cases and scripts automatically, a process known as autohealing. This concept is invaluable when you have large code bases to maintain. In the long run, we want to work to integrate this into the CI flow (CI = Continuous Integration) to achieve more automation and minimize manual effort.

How can generative AI streamline the testing process?

- By quickly creating artifacts such as requirements and test specifications. In a short time, AI can generate test cases based on requirements or code, but also visualize results so that decision-makers have a clear basis for evaluation.

- AI will help take automation to the next level, but manual testing will still be needed in the future. Our mission is to help customers identify where AI adds value. If we can find bugs early, we can keep development costs down - this is where AI can really make a difference.

Quality and reliability

What is the role of prompt engineering and technical know-how?

- Using Generative AI effectively requires clear instructions - prompts - that the AI understands and can execute on. Creating effective value-added prompts requires technical expertise - you need to understand testing at both a strategic and technical level.

- Technical skills are also essential when it comes to source criticism. You need to be able to handle both the tools and the context to get reliable results.

How do you ensure quality and reliability when using AI?

- By having testers who are also AI engineers. You need to understand biases that are built into algorithms, models or introduced by poor training data. We conduct risk analyses and document the consequences of limited data. Through expertise, source criticism and awareness of bias, risks can be managed.

- In addition, our employees are trained in regulations such as the AI Act; how to handle personal data, what can be saved and where. As a company, we have a responsibility to handle AI with care - using AI at work is not the same as at home.

Bright future

How do you see TestDev Powered by AI developing in the future?

- The foundation is to make customers' processes more efficient. Going forward, we can minimize repetitive steps, achieve higher levels of automation, increase quality with data analytics and streamline software flows by introducing AI into the CI chain. We will be able to guide customers to the right processes, routines and tools - and act as a safe hand when introducing AI.

- Today, AI is mainly used for scripting and test generation, but within five years, AI is likely to be seamlessly integrated into customers' build chains and automation flows. We are just at the beginning of the journey," Maria concludes.

Contact Maria!

Maria loves to talk smart approaches to testing and AI. Give her a call or send her an email and she will be more than happy to help you

Tel: +46 76 890 78 72
E-mail: maria.alemyr@infotiv.se

You can also fill in the form and we will contact you as soon as possible.