Business Central
Ina Rogalev
6. May 2026
Many companies currently have a similar feeling when it comes to AI: they do not want to miss out, but they also do not want to make rushed decisions. In new digitalization and software projects in particular, this creates tension. Alongside the classic questions around processes, interfaces, data, and implementation, another one often comes up: Do we also need to consider AI here?
That question is valid. It only becomes problematic when it is answered with a solution too early. Our project experience shows that the mistake is not thinking about AI too early. The mistake is discussing it too vaguely.
In many companies, AI has automatically become part of the discussion as soon as new systems, processes, or software projects are involved. At the same time, it is not enough to simply mention AI as a requirement. There needs to be a clear idea of what it is supposed to do within the process.
If a company says it wants a chatbot, that is not yet a solid requirement. A meaningful use case only emerges once it is clear what exactly that it is supposed to do. Otherwise, the request quickly stays on the level of familiar AI tools, without defining the actual task more precisely.
A viable AI initiative therefore does not begin with the question of which AI should be used, but with the question of which problem in the company is meant to be solved.
A good example would be use cases in which large volumes of information need to be searched, condensed, and made usable for specific questions. In such cases, AI can create real value. At the same time, this kind of example also shows that good AI projects do not happen on the side. Before anything can even be built, it must be worked out carefully what exactly AI is supposed to do in the process, what the benefit looks like, and what requirements arise from that.
The key questions are therefore always the same:
Only then does the technical discussion become worthwhile.
Alongside the functional and technological question, there is a second level in AI projects that is often underestimated: the human one.
AI is not perceived in companies only as a new tool. It also changes expectations, role models, and a sense of security. That is exactly why it often triggers not only interest, but also reservations.
Typical concerns include:
That is why AI in a software project is not only a question of architecture, data, or use cases. It is also a question of Change Management.
Not every AI idea is automatically bad. But not every one is mature enough yet. That is exactly why it makes sense to talk about prerequisites before talking about tools.
These questions help with the assessment:
Before implementation is discussed, it should be clear what AI is supposed to improve: more speed, more precision, less manual effort, or better access to information.
Even if AI is not implemented right away, the system landscape should be designed in a way that allows later scenarios without having to rethink everything from scratch.
AI changes not only processes, but also expectations and routines. That is why possible questions, concerns, or uncertainties within the team should be taken into account early on.
Anyone who clarifies these points in advance creates a much stronger basis for sound decisions. Then AI does not become the next source of pressure in a project, but a building block that provides support where it actually creates value.
STEFAN SCHADE
Senior Project Manager
INA ROGALEV
Junior Marketing Manager
Other articles that might be interesting for you
