Teams that carry transformation: psychological safety is not a comfort topic

The most expensive sentence in a project is: "I saw that coming three months ago."

Almost every failed initiative had people who saw it coming early. An engineer who knew the interface would not hold. A project lead who found the timeline unrealistic from the start. A department head who could tell her team would not follow the new process.

They did not say it — or they said it once, at an unfortunate moment, and never again.

That is not a character problem. It is a system problem. And it is why psychological safety in technical initiatives is not a soft topic but a hard project indicator.

What psychological safety actually means

Psychological safety does not mean harmony, absence of conflict, or everybody always being nice. It simply means people can speak up without fearing punishment, ridicule or being considered incompetent.

In practice you recognise it in unspectacular things:

• Somebody says "I don''t understand this" in a room with senior people.
• A mistake is reported before it is noticed.
• An estimate is revised upwards without anyone having to justify themselves.
• Disagreement comes from people who are not the most senior in the room.

Without it, the number of problems does not change — only their timing. They surface later, in worse shape, with less room to act.

The state of a project can be read from the age of the worst news that has reached leadership.

Leadership in transformation: make decisions visible

For teams, transformation mostly means uncertainty: what still applies? Who decides now? How will my work be measured going forward? Uncertainty cannot be dissolved through motivation, only through clarity.

Three things reliably help:

1. Justify decisions — in writing. A re-readable rationale replaces dozens of follow-up questions and prevents interpretations forming that nobody corrects.
2. Assign [ownership](/en/insights/ownership-statt-rollen) by name. "The team takes care of it" often means nobody does. A named person is not a control mechanism; it relieves everybody else.
3. Name non-decisions. If something is open, saying so is better than letting silence make it look decided.

Async by default — not a tooling question

Asynchronous collaboration is often understood as a tooling topic. In truth it is a question of commitment: when decisions are documented with their reasoning, people can do good work regardless of presence and time zone. Missing a meeting no longer means falling behind.

There is a second effect that is rarely named: async work reduces the advantage of volume. Someone who thinks well but dislikes speaking in groups gets the same access as someone who talks fast. For the quality of technical decisions that matters considerably.

It does not mean there are no conversations. It means conversations are used for what they are good at: negotiating, clarifying, building trust. Not for status reports.

Recruiting: hire for the initiative, not for the job ad

Transformation initiatives often recruit for technology and fail on posture. The requirements list contains tools and years, but not what makes the difference: the ability to work in unclear situations, to justify decisions and to voice uncomfortable truths early.

What we consider more robust:

• Ask about concrete situations instead of self-assessments. Not "how do you handle conflict?" but "tell me about a project where you had to revise an estimate upwards."
• Pay attention to how failure is described. People who can describe setbacks precisely and without blaming have learned from them.
• Communicate realistic expectations. A role that is oversold gets filled twice.
• Take placement seriously: the right person for an initiative matters more than the quickly available one.

Culture is what remains after the third bad day

Culture work is not values on a wall. It is the behaviour of leadership in the moment something goes wrong. Is the search for the person who made the mistake, or for the cause that made it likely?

Teams learn from exactly those moments — permanently. The reaction to the first serious mistake determines how much you get to hear over the following two years.

What follows from this

If you want to check just one thing in a running initiative: ask when the last genuinely uncomfortable piece of news reached leadership, and how it was received.

If such news arrives early and without drama, the initiative will probably hold. If it arrives late or not at all, technology is not the first problem you need to solve.