Get in touch with us
We at accompio will be happy to help you.
20.10.2026
AI can generate code in no time and help development teams turn ideas into functional software faster. But more speed does not automatically mean better software. Without a clear architecture, standardized processes, and an understanding of future operations, systems can quickly become cluttered and difficult to maintain.
Dag Dammann, Senior IT Consultant at accompio, helps companies further develop software development, automation, and IT infrastructure. In an interview, he explains why AI does not replace established architectural principles, what role cloud-native approaches and platform engineering play, and how companies can meaningfully integrate AI into their development processes.
Explains in the full expert interview Dag Dammann, Senior IT Consultant at accompio, how companies can make meaningful use of AI in software development, reduce technical complexity, and make their development processes future-proof.
AI changes software development primarily through speed. Ideas can be translated into program code more quickly, solutions can be prototyped more efficiently, and existing approaches can be reassessed. This creates opportunities to fundamentally rethink processes and technical solutions.
At the same time, the existing challenges do not disappear. Software must continue to be understandable, maintainable, and reliable to operate. As more and more features are developed faster, the effort to test, roll out, and maintain stability in operation also increases.
Here we see a fundamental conflict of objectives: DevOps environmentDevelopers want to provide new features as quickly as possible, while operations must ensure stability and availability. AI can accelerate development – but without appropriate automation and operational processes, the bottleneck can only be shifted.
Companies should therefore not solely measure how quickly code is created. What is crucial is whether the entire value creation process becomes more efficient from the idea through development and deployment to operation.
A central foundation for maintainable software is an architecture with clearly defined components and tasks. Instead of developing a single tool for solving as many problems as possible, individual components should take on specific tasks and be able to be meaningfully combined with each other.
This reduces complexity. Developers need to be able to view fewer relationships at once and can implement changes more purposefully. The same principle applies to AI systems: The clearer a task is defined, the less context the model needs to take into account.
Dag Dammann describes this relationship through the so-called cognitive load. While humans can only process a limited amount of information meaningfully at the same time, AI models operate within a limited context window. Therefore, large, unwieldy systems make both human work and the targeted use of AI more difficult.
Cloud-native architectures and microservices can help to clearly define responsibilities and interfaces. They are not an end in themselves, but they provide a foundation on which applications can be developed, operated, and expanded in a modular manner.
The architecture itself should also not be completely delegated to AI. Dammann deliberately starts with the structure: he determines which components are needed, how they communicate, and which languages or frameworks are used. Subsequently, AI can support the concrete implementation – within comprehensible guidelines.
„The built-in software must be so clear that it can be understood again at some point in the future.“
Dag Dammann, Senior IT Consultant at accompio
Cloud-native approaches help build applications so that they can utilize standardized cloud platforms and orchestration solutions. Containers and platforms like Kubernetes provide established mechanisms that companies do not have to develop anew for each application.
Platform Engineering builds on this idea: Teams receive standardized platforms and reusable workflows to provide and operate applications more efficiently. This makes it easier to integrate automation, security checks, and monitoring into existing processes.
Also standards like Containerization And GitOps creates repeatable processes. This initially causes effort, but later makes operation and collaboration easier. New employees find their way more quickly when they recognize familiar concepts. The same applies to AI: Common technologies and established standards are already well represented in many models. Therefore, they need less to explain individual special solutions and can build on known knowledge.
The advantage therefore lies not only in the technology. Standards create common interfaces, reduce unnecessary complexity, and make it easier to connect development, security, and operation together.
With increasing AI support, it becomes more important to master not just individual programming steps. According to Dammann, so-called T-Shaped DeveloperProfessionals with broad knowledge of the entire software lifecycle and additional depth in a specific area.
This life cycle includes, among other things, planning and architecture, development, testing, release, operation, and monitoring. Those who understand these phases in context can better assess the impact a change has on the overall system.
This perspective is crucial in working with AI as well. Developers must formulate tasks precisely, set appropriate guidelines, and be able to critically evaluate the generated results. This includes understanding technical decisions and analyzing errors in a comprehensible manner, rather than relying solely on the model’s output.
AI thus shifts part of the workload: Less time may need to be allocated to the manual implementation of individual code sections. In return, an understanding of architecture, problem analysis, process knowledge, and the evaluation of technical solutions become more important.
Before scaling up AI-based software development, companies should review and document their existing processes. If processes are unclear or contradictory, automation can even reproduce the underlying problems more quickly and on a larger scale.
A common mistake is to start right away with the biggest and most difficult problem. If the project fails, not only costs are incurred, but there may also be reservations about the new technology.
Dammann’s recommendation therefore states that you should start with a clear goal in mind and initially select a manageable use case. This does not have to solve the company’s biggest problem. What is important is that the approach can be tested and evaluated within a limited framework.
In this way, teams can gain experience, identify technical and organizational challenges, and subsequently gradually expand their use of AI. AI is thus not introduced as an isolated tool, but as part of a well-thought-out development process.
accompio helps companies coordinate development processes, software architecture, and IT infrastructure for their solutions and requirements.
AI offers companies the opportunity to develop software faster and redesign existing processes. However, this advantage becomes sustainable only when architecture, standards, and operating processes adapt.
Modular systems, cloud-native technologies, and standardized processes lay the foundation for ensuring that AI-generated code remains understandable and maintainable. At the same time, development teams need employees who understand the entire software lifecycle, can set clear guidelines, and can assess results professionally.
Dag Dammann's central recommendation: Companies should master the technical fundamentals, make processes transparent, and start with small, controllable steps. Because speed alone is of little value if the direction is not correct.
„If you can’t do something slowly, then you usually can’t do it quickly either.“
Dag Dammann, Senior IT Consultant at accompio
Dag supports companies in integrating AI into software development so that solutions can not only be developed faster, but also be operated and developed in the long term.
AI can generate program code faster, support prototypes, and reevaluate existing solutions. However, architecture, quality control, maintainability, and operation remain important tasks that must continue to be professionally supported.
A clear, modular architecture reduces complexity and separates tasks and responsibilities from each other. This allows developers and AI systems to more easily understand, process, and test individual components.
Cloud Native refers to approaches where applications are designed to operate on cloud platforms and standardized environments. This allows existing platform features, automation, and established operating procedures to be utilized, rather than developing individual solutions for each application.
T-shaped developers possess broad knowledge of the entire software lifecycle while at the same time having in-depth expertise in a specific field. This helps them to target AI and evaluate the results in the overall context.
Companies should first clarify their processes and objectives, document relevant workflows, and start with a manageable use case. After evaluating the results, the deployment can be expanded gradually.
