AI in QA: Roman Kapalo on building an AI-first workflow

AI in QA: Roman Kapalo on building an AI-first workflow

AI is becoming part of everyday work across N-iX, with each role finding its own ways to use it. In our AI in Our Daily Work series, N-iXers share the tools they use, the changes they see in their workflows, and the lessons they have learned from applying AI to real tasks.

The first conversation features Roman Kapalo, Senior AQA Engineer and AQA Community Lead at N-iX. He explains how his team is building an AI-first QA workflow, where AI supports everything from requirements analysis and test planning to automation and final checks.

Roman also shares what pragmatic AI adoption looks like in practice: applying AI to clearly defined problems, measuring its impact, and keeping quality and human verification at the center.

What does your role involve day to day?

We’ve moved beyond separate Manual QA and Automation QA roles toward a General QA Engineer model, where one specialist takes ownership of quality across the entire process.

My work includes requirements analysis, specification reviews, manual and automated test planning, testing, and building automation. I believe this broader ownership is where the industry is heading, and AI is accelerating the shift.

Which AI tools do you use regularly?

Our toolkit has evolved together with our workflow. We started with GitHub Copilot, moved to Cursor, and now use Claude for our daily tasks.

The biggest change is how deeply AI has become integrated into our processes. It now supports different stages of analysis, planning, implementation, and review.

What has changed the most since you started using AI?

We have adopted an AI-first workflow built around specification-driven development and reusable AI skills for recurring tasks.

Every feature now starts with AI-assisted analysis and ends with AI-supported checks. AI has become part of the entire workflow, while engineers remain responsible for decisions and quality.

Can you share a practical example?

One recent game-changer was a provisioning skill that helped us set up environments for test automation. It significantly reduced repetitive setup work and gave the team more time to focus on testing and product quality.

This is pragmatic AI in practice: starting with a clearly defined problem and applying AI where it can create practical value.

How much faster has AI made your work?

In some cases, AI helps us complete tasks several times faster. However, speed alone does not provide the full picture.

Part of the time saved during analysis and development should be reinvested in reviewing, validating, and improving AI-generated output. The real value comes from combining AI’s speed with human expertise.

How do you measure AI’s impact?

We are still developing more precise ways to measure performance, and I would not point to one metric that tells the whole story.

We look at a combination of indicators, including utilization, estimation accuracy, delivery speed, and code quality. Together, they provide a more complete view of AI’s impact on the workflow.

What advice would you give someone using AI in a similar role?

Start by identifying the most repetitive and time-consuming parts of your workflow. These are often the areas where AI can create the clearest immediate value.

Integrate AI into those specific activities, evaluate the results, and adjust the process as your experience grows. Most importantly, always verify the output. Human review remains essential.

Has your perspective on AI changed?

I was initially skeptical, but I wanted to test AI and understand it before forming an opinion.

Having seen its capabilities and how quickly the technology is developing, I now encourage others to start adapting and take an active role in the change. The teams that experiment responsibly and apply AI to real challenges will help shape how it is used across the industry.

AI is already changing both the QA workflow and the role of the QA engineer. The next step is learning how to use it pragmatically, responsibly, and with quality at the center.

Roman’s experience is the first look at how N-iXers are bringing AI into their daily work. Next, we’ll explore other roles, tools, and workflows, with the same focus on real use cases, tangible impact, and lessons worth sharing.

Stay tuned for more practical AI experiences from across N-iX.

About N-iX                                                                                                  

N-iX is a global software development service company that helps businesses across the globe expand their engineering capabilities and develop successful software products. With more than 2,400 professionals and over two decades of experience, N-iX offers expert solutions in software engineering, cloud, data analytics, embedded software, IoT, machine learning, and other tech domains covering a wide variety of sectors, including finance, manufacturing, telecom, supply chain, and retail, to name a few.