by Peter Reddering

In recent years, Data & AI have evolved from an experimental playground to a strategic core component of organizations. Where data science once began as a pioneering domain — driven by curiosity, innovation, and loose experiments — by 2026, the reality has fundamentally changed. Data & AI are no longer a side issue. They impact primary processes, influence decision-making, and, when applied correctly, havesee our third articledirect societal impact.

The shift demands more than better models or faster infrastructure. It requires organizational maturity. It demands clear ownership, robust governance, and conscious choices about how, where, and under what conditions AI is deployed. Successful organizations realize that AI can no longer be a mere afterthought handled by an innovation team or a detached data science group, but must be structurally embedded within the organization.

In this article, we explore this coming-of-age through five propositions and a clear conclusion to move from pioneering to professionalizing.

Thesis 1: As long as ownership and governance remain implicit, Data & AI will remain vulnerable – no matter how advanced the technology.

Many organizations are investing heavily in Data & AI technology. Modern data platforms and powerful AI models: technically, everything seems possible. Yet, we see that the return often lags behind the ambition. Projects stagnate, responsibilities become blurred, and risks only become visible when things go wrong.

The cause often lies not in the technology itself, but almost always in the lack of explicit ownership and governance. Who is responsible for data quality? Who decides if an AI model can go into production? Who intervenes if a model consistently produces undesirable outcomes?

In a pioneering phase, that vagueness is still manageable. Small teams, short lines of communication, room for experimentation. But as soon as Data & AI take on an organization-wide role, implicit ownership becomes a structural risk. Especially in sectors with significant societal impact – such as healthcare, government, and financial services – it is no longer acceptable for responsibilities to lie "somewhere" or with multiple people.

Professional Data & AI organizations make ownership explicit. They appoint data domain owners, model custodians, and decision-making structures. This isn't to slow down AI optimization but rather to make it scalable and reliable. Governance may seem like a bureaucratic layer, but it is a prerequisite for sustainable value creation.

Thesis 2: The gap between business and technology persists without bridge builders like the Data & AI translator.

As long as these worlds aren't effectively connected, misunderstandings will arise. The business expects "smart insights," while the data simply isn't suitable for it, or the exact business question isn't formulated sharply enough. Technical teams build solutions that are correct but barely align with the workflow and information needs. The result: frustration on both sides and solutions that are insufficiently utilized.

This is why mature organizations explicitly invest in bridging roles, such as Data & AI translators. These are not additional project managers, but professionals who understand both the domain and the technology. They translate business questions into analytical capabilities, and technological limitations into realistic expectations. They ensure that what is created is also used effectively and continuously optimized so that solutions do not end up unused on the shelf.

This role is crucial at a time when AI is becoming increasingly complex. Not everyone needs to understand exactly how a model works, but someone needs to be able to explain its potential and impact, where the risks lie, and which choices are being made. They must also be able to define the requirements for the AI solution—such as performance, transparency, and traceability—so that the technical teams know what the solution needs to achieve. Without that translation, AI remains either too technical or too vague—and in both cases, ineffective.

Thesis 3: Without an AI-ready data model and clear architectural principles, scalability remains an illusion.

Many AI initiatives start with a specific use case. A prediction, a classification, a recommendation. That works, at least temporarily. But as soon as organizations want to combine, reuse, or scale multiple use cases, they get stuck. The reason is often that the underlying data model is not prepared for this. Data is fragmented, semantics differ per system, and definitions are implicit or contradictory. What is called "customer" in one dashboard means something else in another. AI models build upon these flaws, thereby amplifying the confusion.

Successful organizations therefore invest in an AI-ready data model: a shared, unambiguous, and understandable description of core concepts and entities that supports analysis and decision-making.

This also includes clear architectural principles and central frameworks, especially now that low-code and no-code solutions are becoming increasingly popular. Without guidelines, dozens of shadow solutions will emerge that deliver value locally but lead to technical and organizational chaos collectively or in structural, long-term use.

Proposition 4: Ethical decision-making is not a checkbox, but a continuous design process.

In the adult Data & AI organization, ethics is shifting from a prerequisite to a core question. Not because legislation enforces it—although regulations like the AI Act play an important role—but also because organizations realize that trust is crucial for adoption. This applies to internal use (users who are more concerned with checking the outcomes of AI than using it) as well as to customers and stakeholders around the organization (what they do, how, why, etc.).

Too often, ethics are reduced to a legal checkbox: does this model comply with the rules, yes or no? But ethical issues are rarely black and white. They concern purpose limitation, explainability, and societal impact. About what is technically possible versus what is desirable. It's important to realize that ethical discussions always precede legislation; by only looking at legislation, an organization will always be behind the curve.

This is why we see progressive organizations explicitly organizing ethical decision-making. They involve multidisciplinary teams, conduct impact analyses, and accept that ethical questions are not "solved" but require continuous attention. Ethics thus becomes not a brake on innovation, but a recognizable steering mechanism for what we want to be able to do with Data and AI.

Proposition 5: BYOAI and shadow IT are not IT problems, but governance issues.

With the rise of generative AI, the use of external tools has increased explosively. Employees are using their own AI assistants, uploading data to online AI services, and automating tasks outside the sight of IT or compliance. This so-called BYOAI (Bring Your Own AI) and shadow IT is understandable; the tools are powerful and easy to use, but they pose a real risk. A ban rarely works. What does work is clear policy: what is allowed, what is not, and why. Professional organizations recognize that shadow IT is a signal of unmet needs. Instead of controlling and punishing, they opt for direction, safe alternatives, and awareness.

Conclusion: Professionalization is not an end point, but takes place based on integration, continuous learning and improvement from culture and a conscious attitude.

The adult Data & AI organization of 2026 is characterized by integration. Cost management (FinOps), security, privacy, compliance, and model governance. These aspects are not seen as separate disciplines, but as an integral part of the data domain.More on this, see our 2nd article)

This calls for cross-departmental collaboration and new skills. But the result is an organization that is not only technically mature but also manageable, explainable, and agile.

The shift from pioneering to professionalizing does not mean that experimentation stops. On the contrary. It means that experiments take place within a framework that balances value, risk, and responsibility.

In 2026, Data & AI have become too important to be optional. Organizations that recognize this and invest in maturity (organizational, ethical, and architectural) are laying the foundation for sustainable impact.

Cmotions helps organizations set direction and map out and execute organizational design issues in a cohesive manner. Data strategy, organization and process design, data management, and its implementation.

Let's brainstorm once about how your organization can grow and make a sustainable impact with data and AI?

Contact the author of this article Peter Reddering to contact Cmotions via info@cmotions.nl or call 088 – 200 75 00 

Want to know how data-mature your company is right now and if you're fully ready for AI? Then take The Data Maturity Scan.