Digital Sovereignty Can’t Happen Without Engineering, Architecture, and Accountability for Data

The discussion about digital sovereignty has entered a new phase. Geopolitical tensions, dependence on global technology platforms, and the rapid development of artificial intelligence have raised a fundamental question: how do companies actually manage their data, where is it processed, and who is accountable for it? Martin Pluschke from Nürnberger Versicherung argues that sovereignty cannot be achieved through strategy papers and high-level statements alone. It requires engineering discipline and scalable architecture.

“What we see in this discussion is a strong combination of high-level strategy and a deep dive into the engineering room,” Pluschke says. In his view, this is the crucial difference between ambition and implementation. Digital sovereignty in the age of AI is not only about formal control over data. It is about an organisation’s ability to understand, structure, protect, and use that data in business processes. “What we need is to get from paperwork to reality, and for that we need real engineering art and scalable architecture,” he says.

The central issue in this transformation is not the AI model itself, but the data on which it is built and applied. Pluschke is direct: companies must first solve the problem that has been in front of them for years. “You have to handle your data. This is what should happen in every company,” he says. Without that, artificial intelligence becomes just another layer placed on top of a fragmented, poorly structured, and often insufficiently understood data foundation.

A particular problem, he argues, is that data is still too often treated as a technical matter rather than a business responsibility. In practice, that is not sustainable. “In terms of data, it is not only a technological thing. It is moreover about data governance,” Pluschke says. This points to one of the major weaknesses in many organisations: when data is discussed, business departments look to IT, while IT waits for the business to define what should be done with the data.

That division of responsibility creates an organisational deadlock. Data is generated in the business, used by the business, and must serve a business purpose. Still, without technological architecture, it cannot become a reliable basis for analytics and AI. Pluschke describes this relationship as a tension between two sides that depend on each other but often fail to take full responsibility. “The biggest problem is that the business does not feel accountable for its own data,” he warns.

Although companies have been talking about data as a new source of value for more than a decade, Pluschke is sharply critical of the actual progress made. He recalls discussions from 2015, when the market was already debating whether data was a goldmine or a landmine. Today, he says, many organisations still have not made the essential shift. “When I look back and when I look now at my organisation and every other organisation, to be honest, nothing has changed,” he says.

What has changed, he adds, is mainly the technology. Tools are now better at handling unstructured data, which matters because much of business information does not live in clean databases but in documents, emails, notes, conversations, and internal systems. Yet the ability to process unstructured data does not solve the underlying issue if organisations still do not know what data they have, where it sits, who owns it, and for what business purpose it is used. “The only thing that has changed is that we can better handle unstructured data in terms of technology,” Pluschke says. “But that we have really sorted out our data, we have not.”

Data silos, therefore, remain one of the biggest barriers. Companies can invest in cloud platforms, data centres, AI systems, and advanced models, but if data remains trapped in departments, applications, and legacy systems, its real value remains limited. Pluschke is frank about the practical challenge: “We still have silos in companies, and this is my hassle every day.”

For that reason, data governance is becoming one of the defining issues of digital transformation. Pluschke does not soften his assessment. “There is no data governance,” he says, describing a situation in which many organisations still need to build real mechanisms of accountability, quality, and data usage. For the future, he believes it is essential that the business side becomes much more involved in understanding data, measuring it, classifying it, and deciding which data is truly needed.

In other words, data competence must move beyond the IT department. Business teams need to understand what is in their data, how it can be used, and where the boundaries of that use are. “You need more competence and more know-how in the business sector to understand what is in the data, how it can be used, and what data is really needed,” Pluschke says. Only then can IT have a true counterpart on the business side, and only then can artificial intelligence rest on a stable foundation for real business applications.