The Stem
Cell

Meridian is a European infrastructure group employing eleven thousand people across seven countries. It builds things that last — bridges, energy grids, water systems — and it has learned to build them well. Its project management methodology is the envy of its sector. Its safety record is exemplary. Its governance structures have been refined over four decades into something approaching a precision instrument.

It is also, by its own admission, exhausted.

Not in the way that comes from overwork, though there is plenty of that. Exhausted in a deeper sense: the kind that sets in when an organization has optimized itself so thoroughly that the space for surprise has been almost completely eliminated. When every decision passes through so many layers of approval that by the time it emerges, no one is quite sure who made it. When the people who joined because they wanted to build something meaningful spend most of their energy navigating the system rather than doing the work.

The Chief Executive, a quiet engineer named Hartmann who had spent thirty years at Meridian before ascending to its leadership, described it to us in our first meeting with characteristic precision.

He had read enough to know that this was not a strategy problem. Not a structure problem. Not, despite what three consecutive consultancies had suggested, a change management problem.

It was a power problem.

Meridian was operating, thoroughly and expensively, in Engineered Power. The processes worked. The coordination worked. The reporting worked. But the living intelligence of the organization — the creativity, the collective judgement, the capacity to sense what was coming before it arrived — had been systematically organized out of existence.

Hartmann wanted to know if it could be invited back.

The Digital Infrastructure Initiative was not chosen because it was the most important project at Meridian. It was chosen because it was the most exposed.

A cross-functional team of fourteen people — engineers, data architects, a procurement specialist, two external partners, and a project director named Yael who had the particular quality of being equally trusted and equally feared by almost everyone she worked with — had been assembled to design the company's next-generation asset monitoring platform. A five-year programme. Three hundred million euros. Significant visibility.

It was also, structurally, a nightmare. The team members reported to seven different division heads. Their budgets came from four different cost centres. Their outputs fed into three different strategic programmes, each owned by a different member of the executive committee. The project had been running for eight months before Confluent was introduced, and in those eight months it had produced, by Yael's own count, forty-seven PowerPoint presentations and approximately zero decisions.

This is what Restricted Power looks like from the inside: not laziness, not incompetence, not bad faith. Fourteen capable people, each of them individually excellent, collectively unable to act because the invisible architecture of the organization — its habits, its hierarchies, its unspoken rules about who is allowed to initiate what — had calcified around them like sediment around a pipe.

Two weeks before Confluent began working with the team, Hartmann did something unusual.

He sent a personal message — not an email, a handwritten note — to each of the fourteen members of the Digital Infrastructure Initiative. The note said, in its entirety:

He then, deliberately and visibly, stopped attending the project's steering committee meetings.

This is what we mean by the silent contract. Not absence — presence of a particular kind. The CEO had not abdicated. He had done something more difficult: he had transferred trust in a system that did not know how to receive it. He had created, in the organizational immune system that would have otherwise treated any anomaly as a threat to be neutralized, a small but crucial zone of tolerance.

He did not tell the team what to do with that zone. He did not set outcomes or define success criteria beyond the technical deliverables already established. He simply said, in effect: the usual rules do not apply here, and I will back you if they try to reimpose themselves.

The first Confluent session with the team lasted four hours. We did not discuss the platform. We did not review the project plan. We did not open a single PowerPoint.

We began with a question: "What do you know about this project that you have not yet been able to say out loud?"

The silence that followed lasted nearly three minutes. In organizational terms, three minutes of silence in a room of fourteen people is an eternity. Most groups break it within thirty seconds with a joke, a procedural comment, or a question about the agenda. This group held it.

What emerged from that silence was not a list of problems. It was something closer to a collective exhale. The procurement specialist said she had known for six months that a key technical assumption in the platform architecture was wrong, but had not found a way to surface it without appearing to undermine the project director. One of the external partners said he had been modifying his own recommendations to match what he sensed the client wanted to hear. Yael said she had been managing upward so intensively that she had largely stopped talking to her own team.

None of this was pathological. All of it was entirely rational given the organizational conditions these people were operating in. But it had created, beneath the surface of the project, a layer of unexpressed truth that was consuming enormous energy and preventing the team from accessing its actual collective intelligence.

Over the following weeks, something began to shift. Not dramatically — systemic change rarely announces itself. But the team started operating differently in ways that were difficult to attribute to any single intervention.

Yael began holding what she called "pneuma moments" at the start of each weekly session — fifteen minutes of silence, or sometimes a short walk outside the building, before any agenda item was opened. Several team members reported that these moments felt absurd at first, and then indispensable.

The data architect, a methodical man named Koenig who had spent twenty years building systems designed to eliminate ambiguity, began asking a question he had never asked before: "What is this data not telling us?" He later described this shift as the most professionally significant of his career — not because it changed what he built, but because it changed what he noticed.

The team developed a practice they called "the honest update" — a standing item in their fortnightly report where one member was asked to say something true about the project that had not appeared in any official document. It was voluntary. It was never formally mandated. Within three months it had become the most anticipated item on the agenda.

It came, as it always does, from an unexpected direction.

Seven months into the programme, a new divisional director — Meridian had reorganized its northern European operations, and three of the team members now reported to someone who had not been part of the original conversation — attended a project review meeting. His name was Brandt. He was competent, results-oriented, and entirely sincere in his conviction that what he was observing was a problem.

His report to the executive committee described the Digital Infrastructure Initiative as "operating outside standard governance protocols," "lacking in structured reporting discipline," and "showing signs of a project culture that prioritizes process exploration over delivery accountability."

He was not wrong about any of the facts. He was wrong about what they meant.

What happened next was the silent contract made visible.

Hartmann did not intervene in the formal review. He let the process run. But he made three phone calls — to the CFO, to the head of internal audit, and to Brandt himself — in which he said, with characteristic brevity, that he was aware of the review, that he had personally authorized the operating model being used, and that he expected it to be allowed to complete its current phase before any governance changes were considered.

He did not explain. He did not justify. He simply made the boundary clear.

The formal review concluded three weeks later with a recommendation for "enhanced reporting" — a face-saving formula that allowed the system to assert its authority without dismantling what had been built. The team implemented the recommendation in the most minimal way possible and continued.

It did not spread the way change programmes spread. There was no rollout. No cascade. No Meridian-wide "ways of working" initiative with a name and a logo and a mandatory e-learning module.

It began with small things. Other project teams started asking whether they could attend the Digital Infrastructure Initiative's "honest update" sessions as observers. The answer was yes, and after the first session the observers rarely remained observers. The "pneuma moment" practice migrated, informally and without instruction, into two other cross-functional teams — one of which had never had any contact with Confluent.

A middle manager in the asset management division, a woman named Sørensen who had been at Meridian for twelve years and had largely concluded that nothing essential would ever change, attended a session at the invitation of a colleague. She later described what she observed as "the first meeting I have attended in this organization where I believed that the people in the room were actually thinking."

She spent the next six months quietly reorganizing her own team's operating model. She did not ask for permission. She did not report it upward. She simply changed the conditions — the structure of the meetings, the quality of the listening, the questions that were considered legitimate — and waited to see what emerged.

Sørensen did not attribute this to Confluent. She barely knew the word. She attributed it to what she called "remembering that we are human beings, not workflow nodes."

Eighteen months after the Digital Infrastructure Initiative began, Hartmann commissioned a systemic assessment of the organization. Not a performance review. Not a culture survey. A genuine inquiry into how power moved through Meridian — where it circulated freely, where it pooled, where it had stopped moving altogether.

The assessment identified seven teams across the organization that were, by varying degrees, operating in what we would recognize as Creative Power. None of them had been formally enrolled in any programme. Five of them had had direct contact with the Digital Infrastructure Initiative team. Two had not.

The platform was delivered. On time, under budget, and — more unusually — with a team that was more capable at the end of the project than at the beginning.