The Visibility Gap in AI Strategy
Darline Giraud

AI adoption often begins through individual and team experimentation before it becomes visible to senior leadership. This creates a gap between formal strategy and the way work is actually changing across an organization. The article argues that leaders need broader, more direct visibility into how AI is being used in practice so they can support effective adoption, identify emerging risks, and build governance around the reality of work rather than an outdated picture of it.

AI adoption often begins long before leadership formally recognizes it. The challenge for organizations is not simply whether AI is being used, but whether leaders can actually see where and how that change is already reshaping work.

As generative AI spreads across organizations, executives are moving quickly to define strategy, governance, infrastructure, and acceptable-use policies.

Yet beneath these formal initiatives, another transformation is already underway.

Employees are experimenting with AI on their own. Teams are building informal workflows. Individuals are finding ways to eliminate repetitive work, improve analysis, and solve problems faster—often without any official mandate.

This creates a growing gap between the organization leaders believe they are managing and the organization that is actually emerging.

That gap is becoming one of the most important leadership challenges of the AI era.

ㅤ

When the Formal Picture and the Real Organization Diverge

Darline Giraud first encountered this kind of gap while managing technology change inside a large international organization.

On paper, a transformation project appeared successful. It had been delivered on time and within budget.

But conversations with the people doing the actual work revealed a different reality.

Morale was deteriorating under constant task-switching. Teams were losing time to coordination and siloed handoffs. There was little space to work collectively on difficult problems or think critically about better solutions.

Both versions of the organization were technically correct.

They simply described different realities.

The same pattern is now appearing around AI.

Senior leaders typically depend on formal systems to understand what is happening inside their organizations: surveys, dashboards, management reports, town halls, and advisors who translate complexity into something actionable.

These systems were designed for a world in which information moved upward through hierarchy, becoming progressively filtered and structured along the way.

AI does not necessarily move like that.

Its adoption often begins with individuals and small teams, then spreads laterally through informal networks, experimentation, and peer-to-peer learning.

Research into generative AI adoption across European software companies has shown that adoption does not always follow a top-down path. In many cases, employees themselves become the starting point.

This does not mean AI creates a visibility problem from scratch.

Rather, it exposes whether an organization already has the ability to detect meaningful change while it is still emerging.

The key question is therefore not whether AI is being adopted.

It is whether leaders can see where and how adoption is happening before it has already reshaped the way people work.

ㅤ

AI Often Arrives from the Edge

By the time an official AI working group is formed, employees may already be far ahead of the formal strategy.

A coordinator may have developed a prompt that reduces a four-hour task to forty minutes.

Two analysts may be sharing an AI-assisted workflow they developed together, without submitting a request or seeking formal approval.

A junior employee may quietly become the person everyone turns to when they want to understand what AI can actually do.

None of this necessarily appears in an executive briefing.

None of it may have passed through the strategy process.

Yet real organizational change is taking place.

Change often begins this way: not at the top, but at the edge of the organization, among the people closest to the work and closest to the problems.

A formal initiative does not always create adoption.

Sometimes it arrives after adoption has already begun.

When designed well, it can provide structure to what is already happening.

When designed poorly, it can interrupt it.

The real question is not whether experimentation exists.

It is whether anyone at the top can see it.

ㅤ

Why Leaders Struggle to See What Is Changing

The difficulty is not simply a matter of competence.

The people responsible for helping executives understand an organization are often highly skilled at synthesis.

They know how to take complexity, identify the signals that matter, and translate them into a format senior leaders can act on.

But that role is becoming harder.

Many boards and senior leadership teams still lack the AI fluency required to engage comfortably with uncertain, incomplete, or technically complex explanations of what is happening inside their organizations.

When leaders are not equipped to sit with that complexity, the people reporting upward can feel pressure to simplify it.

Instead of presenting an ambiguous reality, they may deliver a cleaner, faster, and more reassuring answer.

This is not necessarily a failure of middle management.

It is what can happen when the organization implicitly asks people to protect leaders from complexity rather than help them understand it.

Traditional reporting systems also assume that information flows upward through formal channels.

AI adoption often does not.

It moves sideways.

It appears through individual judgment, informal networks, workarounds, and peer learning.

It does not raise its hand.

It does not automatically generate a report.

As a result, the formal picture may not be wrong.

It may simply be incomplete.

ㅤ

Silence Can Be Rational

There is another reason why important information may never reach senior leadership.

Employees often understand that speaking up about a gap between strategy and reality comes with a cost.

Pointing out that an official process is not working can be interpreted as criticism of someone else’s work, judgment, or authority.

The person who raises the issue carries the interpersonal risk.

The benefit, however, is shared by the organization as a whole.

That imbalance helps explain why employees may choose silence even when their observations could help the company.

The people closest to the work are often best positioned to see where strategy and reality are diverging.

They may also have the least incentive to say so publicly.

This is why leadership briefings can be factually accurate and still fail to capture what is actually happening.

And it is often in that gap that strategy quietly begins to fail.

ㅤ

The Risk of Formalizing the Wrong Reality

The consequences do not always appear immediately.

Leadership continues with the strategy.

The company invests in infrastructure, selects official platforms, formalizes policies, and builds governance around the approach.

But if the original picture was incomplete, what has been formalized may already be outdated.

The employee who built an effective AI workflow may now be required to use an official tool that performs worse.

The analysts who created a valuable process may find there is no place for it within the approved system.

The informal AI expert inside a team may receive no recognition and gradually stop sharing what they know.

At that point, the organization has not merely failed to notice innovation.

It has actively disrupted it.

This is also where governance risk becomes more serious.

AI governance that exists mainly on paper cannot monitor tools nobody logged, access nobody reviewed, or workflows nobody formally documented.

The real risk may therefore not be AI itself.

It may be the invisible layer of AI activity that sits outside leadership’s field of view.

ㅤ

What Leadership Requires Now

The answer is not necessarily another committee or another layer of reporting.

Leaders need a broader set of information sources—and they need to keep widening those sources as the environment changes.

In complex organizations, good judgment alone is not enough.

Senior leaders need more than one line of sight.

This is especially important with AI because adoption is not simply about automating a task.

AI changes how people work, collaborate, make decisions, and increasingly how they interact with technology as part of everyday judgment.

Without visibility into those shifts, governance becomes guesswork.

Rules may become too weak to catch genuine risks, or so restrictive that they eliminate the value they were intended to protect.

Understanding what is actually happening is therefore not something that comes after AI strategy.

It is a prerequisite for building one.

For AI adoption, leaders need what could be called a second channel of visibility.

Not a shadow organization.

Not another bureaucracy.

Something more practical: a deliberate habit of getting closer to the edge of the organization.

ㅤ

Small Habits Can Reveal More Than New Processes

This does not require a large transformation program.

It can begin with simple, recurring practices.

Managers can ask employees in regular one-on-one meetings:

“What did you try with AI this week?”

Teams can hold short peer-led sessions where employees show each other the workflows they are already using.

Organizations can introduce reverse mentoring, with younger or more digitally fluent employees demonstrating tools to senior leaders, while executives provide the organizational context and judgment needed to apply them responsibly.

The aim is not to bypass management or formal governance.

It is to create a more direct view into how work is changing.

The closer leaders are to real experimentation, the more likely they are to identify both value and risk early.

ㅤ

AI Strategy Begins with Seeing Clearly

Leadership has always required knowing when the map no longer matches the terrain.

AI makes that challenge more visible because change can spread through an organization faster than formal structures can document it.

The strongest AI strategies will not begin with the assumption that leadership already understands what is happening.

They will begin by asking where employees are already experimenting, where value is already being created, where risks are already emerging, and which informal practices deserve support rather than replacement.

The leaders who navigate AI well will be those willing to look beyond the official picture.

They will protect the people who identify the gaps first.

They will listen to the employees closest to the work.

And they will build strategy from what is already happening, rather than assuming transformation begins only when leadership announces it.

Because AI is not entering organizations through strategy alone.

It is entering through people.

ㅤ

This article is an abridged adaptation of “The Visibility Gap in AI Strategy” by Darline Giraud, originally published on IE Insights. The original article is available on IE Insights.

https://www.ie.edu/insights/articles/the-visibility-gap-in-ai-strategy/

Written by

diotimes@diokos.com