8 December 2025
An interview with Eva Scheel
Software developer Eva Scheel has been working at OPTIMAL SYSTEMS for eleven years. Most recently, she played a key role in expanding the enaio® microservices using the new Spring Boot Framework. The use of artificial intelligence is now an essential tool in her programming work. In this interview, the 37-year-old describes how she uses AI, how it has changed her day-to-day work and where she sees the biggest risks.
Eva Scheel
Java software developer
What was your first contact with AI in software development?
My first contact with AI in software development was in interaction with IntelliJ IDEA and GitHub Copilot. Some of my colleagues started integrating Copilot on a trial basis to see how well it would support them in everyday tasks. After that, I decided to sign up for a license myself to test it out.
From the very first time I used it, I was surprised at how accurately Copilot often suggests methods – based on class names, comments, or project structure. The context reference is amazingly good, whether Java code, Maven dependency management, or GitLab build pipelines.
It was fun to test different use cases.
Where and in which areas are you currently using AI?
I use AI when creating standard code, such as DTOs, tests, etc., which reduces repetitive tasks.
But I also let it initially suggest documentation, comments or technical descriptions in the language I need. This applies to the documentation in the code, but also to technical documentation. If necessary, I adjust or add content to fix inconsistencies. The AI also provides good approaches for unit tests or refactoring suggestions. It’s not error-free, especially when it comes to more complex processes, but it’s genuinely helpful as a support tool.
Which tools or platforms with AI support do you use most frequently?
I most often use the chat function of GitHub Copilot directly in my development environment. For me, it's the quickest way to get context-related help with Java-specific problems, Spring error messages or writing complex RegEx expressions. The ability to ask questions and get answers directly in the code without having to switch between IDE and browser makes my work easier.
From classification and indexing to intelligent filing:
Benefit from faster processes for managing your documents!
How has AI changed your day-to-day work?
I use Copilot to quickly find solutions to complex problems or framework-specific questions, and also for text generation or phrasing. But also instead of Google searches, especially for standard problems, syntax, error analysis or explanation of concepts. It saves time. AI is no substitute for my expertise, but it's a powerful tool in everyday life. I work more productively and also more creatively.
What advantages do you see in the use of AI in software development in general?
The use of AI has made my day-to-day work as a developer more efficient. AI saves a lot of time, especially with repetitive tasks. I also see advantages in the fast generation of code examples, documentation, or unit tests – something that's a red rag for some developers for various reasons, but is implemented very well by the AI. Instead of struggling with wording, I often get useful suggestions right in the context of what I’m working on. This also improves the content of my work.
What challenges or problems have you encountered?
I think the biggest disadvantage is that AI is not infallible. The suggestions may seem plausible, but are occasionally wrong, unclear or not optimal, especially when it comes to complex logic or performance considerations. Therefore, you shouldn't blindly rely on the answers generated, but must always check them critically.
Another problem I could see is the potential loss of technical understanding if you rely too heavily on AI. In my opinion, if you don't understand generated solutions and still use them, you can cause more problems than you solve.
"If AI relieves me of routine work, then I have more time for the really exciting tasks."
Eva Scheel
Where do you see the limits or risks of using AI?
An AI like Copilot lacks a deeper understanding of the context and architectural decisions. A generated code snippet may be syntactically correct, but AI cannot currently reliably assess whether it integrates cleanly into existing systems, fulfills security requirements or takes performance aspects into account. I see a further risk in the quality of the training data. If the AI has been trained with outdated or dubious source code, this can lead to security gaps or faulty code snippets.
AI can already provide valuable assistance in many areas of programming work – but it cannot replace a software developer.
How do you personally feel about AI increasingly taking on creative or analytical tasks?
I take a pragmatic view: If AI relieves me of routine work, then I have more time for the really exciting tasks. Perhaps also for tasks for which there would otherwise have been little or no time because they are neglected in daily business. For me, creativity in the backend isn't in the syntax of a programming language, but in architecture, scalability, and system integration. These decisions depend heavily on the business context, but also on priorities. The AI cannot take these factors into account when making a decision.
What responsibility do developers bear when using AI systems?
As developers, we are responsible for how and where AI is used. AI-generated results must be checked, tested and documented in the same way as code written by a human. But data protection is also an important issue. Especially in the backend, where sensitive data is often handled, we have to ensure that this data is protected.
How do you think AI will change software development in the next five to ten years?
I think that AI will continue to change the way we develop software, but not in such a way that developers will become superfluous. The focus of programming will probably be on the right prompt. So we will learn to describe what we want, and the AI will generate initial code suggestions, tests, and documentation. I see my role more as an architect and reviewer – assessing quality and security, and making the key architectural decisions.
I also think that AI will further develop the automation of infrastructure, monitoring, and testing. Systems will increasingly optimize themselves, for example through adaptive scaling in real time.
General Remarks
Software & Services
LOCATIONS