What AI Can’t Decide

Why organisational values are becoming a strategic asset in an age of abundant data

By Cosmin Galu, Leadership Empowerment Trainer · Leadership Land — The Atelier Brief

We have never had more data, dashboards, analyses and predictions than we have today. AI produces information at a speed that would have been hard to imagine a few years ago. Experts inside organizations see real, measured, verifiable data every day: operational results, customer feedback, quality indicators, commercial performance.

And yet, in many companies, trust between leadership and experts feels more fragile than ever.

Why?

Because organizations are not struggling with a shortage of facts. They are struggling with the abundance of them.

This is the paradox of our moment. AI has solved the problem most companies spent two decades trying to solve. It has made information available, fast and cheap. What it has not solved — what it cannot solve — is the question of what to do when the information points in more than one direction.

Picture two people inside the same company: a technical expert and a chief executive. The expert works with precise data. They have measured it, verified it, and can defend their conclusions with clear evidence. The leader operates in a different space. Alongside the same data, they have to weigh the market, the competition, regulation, the cost of capital, future risk, and shifting customer behaviour.

Both see a real part of the situation. Both are right. And that is exactly why the conflict appears.

I am writing this because I encounter this situation frequently. I see it in my own team, in the organizations we work with, and even in the development programmes we design. I see it whenever we try to connect the vision of senior management with the pain points our participants bring into the room. I don’t think there is anything new here. This tension between the operational and the strategic has always existed, and it is a natural part of organizational life.

What has changed is the intensity. In Glassdoor’s Worklife Trends 2026 report, mentions of “misalignment” in employee reviews referring to senior leadership rose by 149% between 2024 and 2025. In the same reviews, “disconnect” was up 24% and “distrust” up 26%. And the trend has continued: at Glassdoor’s mid-2026 check-in, mentions of misalignment were up another 95% compared with the previous year. The tension is not new. The pressure it operates under is.

For the expert, information must be validated before it can be used. For the leader, the decision has to be made before all the information is available.

Think of a strategic investment. The production director would prefer another three months of validation and testing. The CEO has to decide next week. Neither of them is wrong. They are simply answering to different constraints.

The problem arises when each assumes the other sees the same reality.

How distrust is built

In organizations, distrust rarely begins with bad intentions. Most of the time, it begins with a gap in perspective.

The expert asks: “If the data is this clear, why was this decision made?”

The leader asks: “How do I explain that this data is one of five variables I had to weigh?”

Between those two questions, frustration is born. Then emotion. Then distrust.

People are not afraid of difficult decisions. They are afraid of decisions they cannot understand and cannot anticipate.

You can tell fairly easily whether this problem exists in your organization:

Decisions are contested after they have been made.
Experts frequently say that something “makes no sense”.
People ask for additional justification for every change.
Conversations start sounding like “us versus them”.
Different teams operate with different versions of the truth.
Managers spend increasing amounts of time defending decisions already taken.

These signals do not necessarily point to a competence problem or a communication problem. More often, they point to a trust problem.

And here is where an uncomfortable truth appears for many leaders.

What cannot be aligned, and what can

We cannot align people at the level of all the facts.

We do not have access to the same information. We do not carry the same responsibilities. We do not manage the same risks. We do not see the same reality from the same position.

We can, however, align people at the level of the values that guide decisions.

Values have an advantage that facts have lost. They are stable in a world of constant change. Data changes. Priorities change. Markets change. Values should stay constant.

More than that, the real values of an organization have already passed the test of reality. They are not the ones written on the walls or in the slide deck. They are the behaviours that have contributed to success repeatedly. They are the things people recognise instantly and describe by saying:

Yes, that’s how we work when we work well.

An expert may not have access to all the information behind a decision. But they can observe whether that decision respects the principles the organization has demonstrated over time.

The key word is “demonstrated”. Not declared. Demonstrated.

This is why I believe alignment on values is one of the few forms of alignment that survives disagreement about the facts.

When people cannot verify all the data behind a decision, they look for something else. They look for evidence that the decision was made in line with the organization’s principles. That is where trust begins to be built.

And this is precisely the territory AI does not enter. A model can tell you what the data shows. It can even tell you what the most probable outcome is. It cannot tell your organization what it is willing to trade away, what it will not compromise on, or which principle takes precedence when two legitimate priorities collide. Those are not calculations. They are commitments.

Five behaviours that can be trained

Values are not a branding exercise. They are a set of observable behaviours. And behaviours, unlike intentions, can be practised.

1. A leader who builds trust makes the decision-making process visible.

Not just the final decision, but the path to it. They explain what information was considered, what risks were assessed, and what remained outside their control.

From the roomWe simulate exactly this. A team builds a structure together, but each person receives a constraint the others don’t know about. Within minutes, half the group is convinced someone is sabotaging the work. Nobody is. Everyone is respecting their own data. The conversation that follows changes how people communicate decisions afterwards.

2. A leader who builds trust validates expertise without surrendering responsibility for the decision.

They don’t say “your data is wrong”. They say “your data is correct, and it is one of the important variables in this decision”.

From the roomWe train this through an exercise where participants first argue their own position, then argue the opposing one convincingly. Acknowledging the other side’s strongest point — before any counterargument — becomes a reflex rather than an effort.

3. A leader who builds trust involves experts before the decision, not after.

People accept a conclusion that differs from their recommendation far more easily when they know their perspective was heard and weighed.

From the roomThe tool we practise here is a single question, asked at the right moment, that moves ownership of the route to the expert while keeping ownership of the destination with the leader. At follow-up sessions, participants name it the most frequently used tool from the entire programme.

4. A leader who builds trust communicates the reasoning, not just the conclusion.

A decision without explanation is received as an order. A decision accompanied by the thinking behind it becomes an invitation to understand.

From the roomWe work with a four-step communication structure that starts from the decision and builds the argument backwards, so the core message reaches the whole audience rather than only those who listened to the end. It is practised on participants’ real presentations.

5. A leader who builds trust makes the guiding value explicit.

They say: “We chose this option because we prioritise customer safety, exactly as we did in similar situations before.” At that moment, people are given a trust criterion they can verify for themselves.

From the roomIn values alignment workshops, the principles are not invented on a flipchart. They are extracted from the experiences where the team worked at its best, then formulated as recognisable decision rules: short, positive, verifiable in behaviour.

A question that changes the direction

The next time a serious disagreement surfaces in your team, I’d suggest not asking: “Who is right?” Ask instead:

Which values are we deciding by?

The first question produces camps. The second produces alignment.

In high-performing organizations, alignment does not mean everyone agrees. It means people understand how a decision was made and can verify that it respects the principles the organization claims to operate by.

None of this is new. Aristotle described the same mechanism 2,400 years ago: when an audience cannot verify the argument itself — what he called logos — they turn to the demonstrated character of the person making it, the ethos. The ethos of an organization is its demonstrated values. The expert who cannot verify the data behind a decision verifies, instead, the character of that decision.

Which is, in fact, the best confirmation of the idea: a principle that has held for two and a half millennia has exactly the properties we attribute to values. It is constant, it is validated by experience, and it is easy to recognise. We did not invent it. We have only tried to turn it into behaviours that can be trained.

In a world where data multiplies faster than our capacity to interpret it, trust becomes a competitive advantage. And trust is not built through access to all the information. It is built through consistency between decisions and demonstrated values.

Perhaps this is one of the most important leadership lessons of the AI era. As facts become more numerous and certainty becomes rarer, the differentiator is no longer who has the data. Everyone has the data. The differentiator is whether people can trust how you decide when the data doesn’t decide for you.

That is the one thing AI cannot replace. And it happens to be the one thing you can build deliberately. ■

Leadership Land works with management teams in regulated industries — pharma, automotive, heavy industry — turning values alignment from a statement into trained behaviour.