From Writing Code to Orchestration: How AI is changing Software Engineering at SupplyOn

Artificial intelligence is fundamentally transforming software development. But what does this mean in practice for developers in their day-to-day work? At SupplyOn, our focus has moved far beyond simple code completion in the editor. Teams here are actively shaping the transition to modern, agent-based workflows.

In this interview, Falco (Senior Software Engineer) and János (Lead Engineer) discuss the shift from manual syntax writing to architectural thinking, the courage to prototype, and why human judgment is more crucial than ever.

If we were to look over your shoulder at work today: What has fundamentally changed compared to the past?

János: A couple of years ago, software engineering was still traditional craftwork. You typed syntax line by line yourself or manually copied recurring code snippets. Then came tab completions – much like predictive text on a smartphone suggesting the next word to insert with a tap – which noticeably accelerated repetitive tasks. Today, intelligent agents handle a large part of the time-consuming routine tasks in code generation. This creates valuable space to focus on what matters most. In my role today, rather than simply typing lines of code, the focus is on precise prompting, orchestrating agents, and conducting thorough architecture and code reviews to ensure the highest quality standards.

Falco: Troubleshooting and debugging have also become significantly more efficient today. In the past, developers often spent hours tracking down individual typos or syntax errors. Today, AI provides targeted support in error analysis: I ask the agent where the issue lies in the code, and I immediately receive sound pointers. This allows developers to focus much more strongly on what the software architecture needs to accomplish and how it best fulfills our customers’ business requirements.

Neither of you settled for the status quo: How did the development of an internal coding agent and the first AI-native team come about?

Falco: I noticed that developers often encountered the exact same bottleneck when using AI. You enter a prompt, wait for feedback, review it, and prompt again. Constantly switching back and forth takes a lot of mental energy and interrupts your workflow. I wanted to see how we could automate recurring steps even further so processes could run smoothly in parallel. Inspired by SupplyOn’s clear strategic focus on AI, I took advantage of internal leeway to independently build a prototype for an internal development agent.

János: In my case, it was about setting up a high-performing team for our new AI-native sourcing solution. The exciting part was that the biggest levers initially had nothing to do with the models themselves, but rather with classic leadership and process questions. How do you form a cohesive unit? How do we establish transparent best practices around AI tools? AI provides an innovative working environment, but close collaboration, open communication, and mutual trust remain the foundation for shared success.

AI generates code remarkably fast. What skills does a modern software engineer really need today?

János: AI agents aren’t just fast, they are also very eager to generate vast amounts of code. That is precisely where engineering experience and the human in the loop become indispensable. It is not just about getting a feature to work quickly today, it is about how that software will hold up in one or two years. Developers have to establish clear architectural boundaries, guide the AI tools, and continuously clean up and streamline the codebase so it remains lean and maintainable over time. In global supply chains, our customers rely on truly resilient software and taking responsibility for that long-term quality and stability always rests with people.

Falco: In the past, the main task often involved breaking a user story down into small, individual tasks and coding classes or methods. Today, we view the system as a whole. How does a feature integrate seamlessly into the overall solution? How do we create a consistently frictionless user experience? For modern software engineers, I see three essential profiles emerging:

• the Rapid Prototyper, who tests new technologies and ideas quickly,
• the Product Builder, who turns them into market-ready, scalable products, and
• the Maintainer, who ensures long-term stability, security, and ongoing evolution in live operations.

To explore such new roles and build prototypes independently, you need the right backing from the organization: What kind of environment and working conditions did you encounter when pursuing these initiatives?

Falco: The support was there right from the start, both from my team and my department. When the strategic impulse came to see what was possible with AI, I simply started prototyping. Whenever I presented my ideas, I was met with open arms, nobody told me to stop. The key is balance: your deliverables within the regular scope must be reliably met. As long as that is covered, you are given the latitude to test new things. That applies beyond engineering as well: if you want to join cross-functional initiatives or leadership topics, teams readily include you and support that time.

János: I experienced that trust firsthand: after just two months, I was given the opportunity to step into team leadership. Receiving that level of trust so early on is exceptional. Because our team was newly formed, I actively reached out across different areas at SupplyOn to validate our direction, which is also how discussions with Falco began. Whether discussing architecture or organizational topics, everyone has been remarkably open and supportive. That cross-functional collaboration provides the foundation for building a strong team culture.

If someone new joins your team tomorrow: What can that person look back on with pride after six months?

János: New colleagues jump right into cutting-edge, agent-based development environments and work on real AI solutions. Thanks to these toolchains, onboarding happens fast: new team members quickly take on active responsibility and help shape the product. Because technology evolves so dynamically, you constantly need to stay on your toes. Looking back after six months, you see that you have truly contributed at the very forefront of modern product development.

Falco: From day one, new team members contribute directly to market-relevant applications that deliver real customer value. The team actively supports newcomers in finding their footing in this new development environment. You get to experience firsthand how features you helped build go live with real customers. Anyone bringing curiosity, systems thinking, and enthusiasm for modern software engineering will find SupplyOn the ideal place to make an early impact.