Artificial intelligence may be changing how businesses operate, but its success depends on something far more fundamental: the quality of the data behind it. Without reliable data and effective data management, even the most advanced AI solutions can struggle to deliver meaningful results. So, what does it actually take to manage data well, and why should organizations care? We spoke with Prof. Dr. Bas van Gils, Managing Partner at Strategy Alliance and Professor at Antwerp Management School, to explore the connection between data management, data quality and AI.
Most organizations have more data than they know what to do with. Customer information, financial records, operational systems, analytics platforms and countless documents all contribute to a growing collection of information.
But simply having data doesn't make an organization data-driven.
According to Bas, the real value of data comes from how it is managed throughout its lifecycle, and ultimately, how it is used.
Organizations need to understand what data they require, how it should be structured, where it belongs and how it can be maintained over time. They also need the right foundations for security, privacy and risk management, supported by clear governance.
Because data, on its own, doesn't create value.
As Bas explains,
“Data is just data. That's not going to help you. It only gives you value when you actually use it.”
This is where data management becomes more than a technical responsibility. It becomes an important part of how an organization operates, makes decisions and develops new capabilities.
Becoming more data-driven is an ambition shared by organizations across industries. But what does that ambition look like in practice?
Bas regularly encounters this question in his work as a consultant and educator.
He describes how organizations often express a desire to become more data-centric, only to struggle when asked what that actually means.
And that is part of the challenge.
Being data-driven isn't simply about collecting more information, introducing new dashboards or investing in advanced technology. It requires a shared understanding of data and how it should be managed.
Consider something as ordinary as the definition of a customer.
For a sales or marketing department, a customer might include anyone the organization could potentially sell to.
For finance, a customer may only be someone who has already made a purchase.
Both perspectives can make sense within their respective functions. But what happens when the organization needs a single, reliable customer list?
Suddenly, a seemingly simple request becomes a question of definitions, ownership and consistency.
This is where data governance, metadata management and data quality become essential.
The challenge isn't always that the information is incorrect. Sometimes, different parts of the business simply mean different things when they use the same words.
Without common definitions and a shared framework, even basic reporting and analysis can become unnecessarily complicated.
As AI becomes more integrated into business operations, the importance of data management is becoming increasingly difficult to ignore.
And Bas makes the connection through a practical example from his own consultancy experience.
While working with a drinking water company, he encountered a use case involving AI and water quality monitoring.
Traditional chemical testing can take considerable time. Biological indicators, on the other hand, can react much more quickly to changes in water quality.
By using data about water characteristics and known quality assessments, AI models can be trained to recognize patterns associated with safe or unsafe drinking water.
The process illustrates a fundamental principle of machine learning.
But what happens when the underlying data is unreliable, incomplete or poorly understood?
Even a sophisticated algorithm has limitations if the data used to develop and evaluate it is not fit for purpose.
Bas puts it simply:
“We need good data management that have good data.”
His reasoning follows a clear chain: effective data management supports good data, which in turn creates better conditions for AI to deliver value.
And no amount of advanced technology can fully compensate for neglecting those foundations.
When people hear the term data quality, they often think about correcting errors, removing duplicates or ensuring that information is complete.
Those activities matter. But managing data effectively involves a much broader set of responsibilities.
Who defines what information means?
Who is responsible for maintaining it?
How are disagreements resolved?
And who has the authority to make decisions about an organization's data assets?
These are questions of data governance.
Bas describes governance in practical terms as establishing decision rights, policies and controls for how data should be managed.
It is about knowing who can make a decision, who is responsible when something needs to be corrected and how the organization should respond when different needs or interpretations conflict.
And these questions reach well beyond the IT department.
Business analysts, architects, data scientists, data owners and other professionals all contribute to how data is understood and used.
For organizations working toward greater data maturity, developing a common language across these roles is an important starting point.
With so many disciplines involved in data management, how can organizations build a common understanding?
One widely used foundation is the DAMA Data Management Body of Knowledge (DAMA-DMBOK), which provides a structured approach to the different areas of data management.
The framework covers topics such as data architecture, data modeling, data security, integration, metadata, data quality and data governance.
Rather than treating these as isolated technical disciplines, it helps professionals understand how they fit together.
Bas explains data management through the lifecycle of data: planning and designing it, enabling and maintaining it, and ultimately using it to create value.
To make this possible, organizations need both a solid foundation and governance that connects the different activities.
Understanding these relationships helps turn a collection of technical concepts into something much more useful: a practical way of thinking about data as an organizational asset.
Understanding the principles is one thing. Applying them in a real organization is another.
This is why Bas emphasizes the importance of combining academic theory with practical examples when teaching data management.
As an active consultant, professor and author of Data Management: A Gentle Introduction, he brings experience from both research and real-world projects into the classroom.
His approach follows three steps.
The aim is not simply to memorize a textbook, but to understand the concepts well enough to recognize their relevance in practice.
As organizations continue to invest in AI, analytics and digital transformation, the underlying questions remain remarkably consistent.
Do we understand our data?
Can we trust it?
Do we have clear responsibilities for managing it?
And are we using it in ways that create meaningful value?
Technology will continue to evolve. But the need for reliable data, shared definitions and effective governance is not going away.
And perhaps that is the most important lesson: before asking what AI can do for your organization, it is worth asking whether your data is ready to support it.
Want to build a stronger foundation in data management? Explore Glasspaper's Data Fundamentals: Certified Data Management Program (CDMP) with Prof. Dr. Bas van Gils and develop a structured understanding of data management, data quality and governance.
