When Cross-Functional Teams Fail: A Structural Lesson for AI Governance in Europe

Recently I read an insightful article in Forbes discussing why cross-functional teams often fail in organisations.

Original article:
https://www.forbes.com/sites/tracybrower/2025/09/08/why-cross-functiona…

The article highlights a common organisational reality:
many organisations consist of highly capable functional units — marketing, engineering, finance, operations — yet collaboration across those units frequently breaks down.

The reason is rarely a lack of talent.
More often, it is the absence of shared clarity.

Different departments operate with different assumptions about:

what the problem actually is

what success looks like

who owns which decisions

why certain trade-offs are being made


When these elements remain implicit, cross-functional cooperation becomes fragile and slow.

This observation may appear to concern only corporate management.
However, it also reflects a structural challenge that Europe currently faces in AI governance.

Europe operates in an environment with:

27 Member States

24 official languages

multiple supervisory authorities

diverse legal traditions and policy priorities


In many ways, this resembles a very large cross-functional organisation.

When policies, interpretations, and enforcement approaches differ across jurisdictions, implementation friction inevitably appears.
Even well-designed regulations can experience divergence once they meet operational reality.

From my experience in enterprise governance and risk management, successful coordination typically requires several elements to be made explicit:

Concept
What exactly are we talking about?

Intent
What outcome are we trying to achieve?

Boundary
Who is responsible for which decisions?

Rationale
Why was this decision taken?

When these elements remain implicit, organisations rely on informal interpretation.
Over time this leads to drift — both semantic and operational.

When they are explicitly documented and shared, collaboration becomes significantly more stable, even across large and diverse organisations.


---

Addressing the Language Challenge

Europe’s governance environment also includes a structural factor that few corporate organisations face: 24 official languages.

In multilingual environments, translation is not merely a linguistic task.
It can gradually alter the interpretation of policy concepts, responsibilities, and decision logic.

Even when translations are accurate, subtle shifts in meaning may appear when terms travel across languages and administrative contexts.

One practical way to reduce this semantic drift is to anchor governance discussions around a small set of shared structural elements — for example:

Concept — the core policy object being addressed

Intent — the outcome the policy seeks to achieve

Boundary — the scope of authority or responsibility

Rationale — the reasoning that supports the decision


When these elements are explicitly structured, translations can reference the same underlying governance structure rather than relying solely on textual interpretation.

In effect, such a structure can function as a shared protocol across languages, helping preserve meaning even when policies are discussed, interpreted, and implemented in multiple linguistic contexts.

This does not eliminate the richness of Europe’s linguistic diversity.
Instead, it provides a stable reference layer that helps different actors maintain alignment while working in their own languages.


---

In corporate environments, this type of structured clarity has long been used to coordinate complex operations.
A similar approach may also help support consistent implementation of AI governance frameworks across multiple jurisdictions.

Importantly, such structures do not restrict innovation.
Instead, they allow innovation and governance to evolve together by making assumptions transparent and decisions traceable.

In other words, governance becomes not only a mechanism of oversight but also an infrastructure for coordination.

Europe has already demonstrated global leadership by establishing a comprehensive regulatory framework for AI.

The next challenge is ensuring that this framework operates effectively in real-world implementation across many languages, institutions, and operational contexts.

The experience of cross-functional organisations suggests that success often depends less on formal authority and more on the clarity of shared structures that guide collaboration.

If those structures are designed well, complexity can become manageable — and even a source of strength.

Tagi
Member States ai regulation discussion blog public sector Best Practice