Data-driven work is constantly evolving. One future possibility is working with AI agents instead of dashboards. At the Port of Antwerp-Bruges (POAB), this approach is already being successfully implemented.

POAB operates under the principle of “No AI without a solid foundation”: First, structure, quality, and ownership; only then advanced applications such as machine learning and AI agents.

During the inspiration session on April 21, Inge Lemmens (POAB) and Johan Saton (Senior Data Management Consultant at Valid) shared insights from their experience at the Port of Antwerp-Bruges (POAB), explaining how data-driven work and the transition to AI only truly deliver value when the fundamentals are in place.

The session highlighted how AI applications are used in everyday practice and why they only work reliably when data, governance, and ownership are properly structured.

Start with the basics, not with AI

AI can only be used reliably and responsibly when data is accessible, consistent, and of high quality. At the Port of Antwerp-Bruges, therefore, the decision was made not to start with AI applications, but rather to begin by getting the data and information in order. This meant first answering some fundamental questions:

• What data do we have?
• Where is this data located?
• What do the data mean?
• Who is responsible for the data?
• What is the data lifecycle?
• How is the data classified?
• How is data quality ensured?

Only once these questions have been answered will there be an opportunity to use data systematically to gain insights, drive decision-making, and foster innovation.

Bring order to your data with governance

A solid data foundation starts with data governance—not as a theoretical model, but as a practical framework for collaboration within the organization. At POAB, data governance is structured around clear roles such as data owners, data users, and a data governance team that facilitates and connects. Together, they establish agreements on definitions, ownership, quality, privacy, and security. Data governance affects the entire organization and requires a phased approach. By starting small and expanding step by step, cohesion is created without it becoming a cumbersome or bureaucratic process.

Make data understandable and reliable

An important part of the approach is documenting semantics and metadata. After all, data only has value if everyone understands terms such as “inbound journey,” “berth,” or “transit time” in the same way. By documenting data in a data catalog with clear definitions, relationships, and quality standards, the data becomes:

• easier to find
• easier to understand
• more reusable

This forms the foundation for reliable dashboards, reports, decision-making, and AI.

AI as the next step, not the starting point

During the session, it became clear that AI is not a standalone solution, but a logical next step. At the same time, you can already apply AI in a targeted manner to areas of data that are already in order, while the groundwork for other domains is still being laid. Without data infrastructure, governance, and BI, an organization-wide AI initiative remains vulnerable.

POAB therefore adheres to the principle: no AI without a solid foundation. First come structure, quality, and ownership; only then do we move on to advanced applications such as machine learning and AI agents. This ensures that AI is used in a sustainable manner and builds trust in the results.

People make the difference

Technology alone is not enough. Successful data-driven work requires engaged employees, clear communication, and ongoing stakeholder management. Change rarely follows a linear path: it’s often difficult at the start, messy in the middle, and rewarding at the end. Patience, understanding, and collaboration are crucial in this process. Only when people understand why changes are necessary and how they themselves can contribute will a lasting impact be achieved.