Organizational Entropy: What Physics Teaches Us About The Fate Of Companies

In 1944 the physicist Erwin Schrödinger asked a simple question: how do living organisms maintain their order in a universe that tends towards chaos? Eighty years later, that question has never been more relevant to the people who run companies.

The Invisible Disorder

You know the feeling. Monday morning, 8.30, executive committee. Your CFO presents some numbers. Your HR director presents others. They do not match. You spend the next hour not deciding, but working out why the data diverges. Reconciling parallel truths. Doing, without knowing it, the work of a system that does not exist.

That scene repeats itself in every organization in the world. And it has a scientific name: entropy.

In physics, entropy measures the disorder of a system. The second law of thermodynamics states that in any isolated system, entropy can only increase. The ice cube melts. The coffee cools. Order tends towards disorder. It is inevitable.

Organizations obey the same law.

The Second Law Applied To Companies

Watch any company over a decade. At the start, when the team fits in one room, everything is clear. Everyone knows what everyone else does. Information circulates naturally. Decisions are fast.

Then the company grows. Departments are created. Systems stack up. SAP for finance. Workday for HR. Jira for projects. Salesforce for sales. Each system becomes a silo. Each silo develops its own truth.

The 2019 org chart no longer reflects the reality of 2024. The job descriptions describe roles that no longer exist. The documented processes are bypassed daily. The decisions taken two years ago — and the reasons for them — are forgotten.

That is organizational entropy. It is not the result of bad management. It is the result of time passing. It is the natural tendency of any complex system towards disorder.

And as in physics, it can only increase. Unless something intervenes.

The Hidden Cost Of Chaos

We have tried to measure that entropy. In a company of 5,000 people, our observations suggest that 70 to 80% of senior executives’ time goes not into deciding, but into reconciling information. Asking several people for the same data. Working out why it differs. Mentally reconstructing the links between projects, teams, and objectives. Confirming that everyone is in fact talking about the same thing.

That is not management. That is informational plumbing.

Translated into euros, it represents several million a year in wasted managerial time. Not counting the sub-optimal decisions taken on incomplete data. Not counting the exhaustion of leaders who carry, inside their heads, the weight of a human middleware.

The most troubling part: that cost is invisible. It appears on no dashboard. It is diluted into the everyday. We call it “complexity”. We accept it as a given.

The Traditional Remedies And Their Limits

Faced with this growing disorder, organizations have developed defence mechanisms. Meetings to synchronize information. Reporting to create visibility. Processes to structure interactions. Consultants to diagnose and reorganize.

Those mechanisms have one thing in common: they inject order at a point in time. A reorganization puts things back in place. A consultant produces a clear map. A new process structures the exchanges.

But entropy always regains the upper hand. Six months after the reorganization, the real structure has already diverged from the official one. The consultant’s diagnosis is obsolete before it is implemented. The process is bypassed as soon as it becomes constraining.

It is like bailing out a boat that is taking on water. The effort is constant. The result is temporary. And the crew wears out.

Negentropy: Schrödinger’s Intuition

In 1944, in his book “What is Life?”, Erwin Schrödinger — the same man who imagined the famous cat both dead and alive — took on a fundamental question: how do living organisms maintain their internal order when everything in the universe tends towards disorder?

His answer came down to one word: negentropy. Literally, negative entropy.

A living organism, Schrödinger explained, is not an isolated system. It exchanges constantly with its environment. It imports order — as food, as energy — and exports disorder — as heat, as waste. That permanent metabolism lets it maintain its internal structure. Resist degradation. Stay alive.

Schrödinger’s insight was revolutionary: life itself is a negentropic process. A permanent fight against disorder, made possible by a continuous flow of structuring energy.

Transposing Negentropy To Organizations

Can that intuition be applied to organizations?

That is the question we asked ourselves while designing LENSX. Not as a commercial product — at least, not initially — but as an experiment. An attempt to create a negentropic system for organizations.

The hypothesis was simple: if organizational entropy comes from the fragmentation of information, the breaking of links, the loss of memory, then a system that unifies, connects, and preserves ought mechanically to reduce that entropy.

Not at a point in time, like a reorganization. Continuously, like a metabolism.

The Architecture Of Negentropy

The design of LENSX rests on an architectural principle it took us years to formulate: organizational data is not two-dimensional. It is three-dimensional.

In a classical system — a relational database, an ERP — data is stored in tables. Rows and columns. To connect two elements you have to create a junction table. A mapping. An integration.

That is why companies struggle so much to hold a unified view. Each system has its own structure. Connecting them takes considerable effort. And that effort has to be renewed at every change.

Our approach is different. Each object — a position, a project, a KPI, a person — carries its own connections within itself. Metadata describing its links to other objects. Not in an external table, but in the object itself.

That is a third dimension. A relational dimension.

The result: the company’s fragmented data — whether it comes from SAP, from Workday, from Excel files — enters the system and finds itself automatically connected. Not by a six-month integration project, but by the structure of the system itself.

A KPI is linked to the actions that affect it. An action is linked to the people responsible for it. A person is linked to their skills. A skill is linked to the training available.

Click on any element and you see its complete context instantly. What influences it. What it influences. Who is responsible. What is blocking.

In seconds, not after weeks of meetings.

Artificial Intelligence As An Accelerator

This is where artificial intelligence comes in. Not as a marketing gadget — “we do AI too” — but as an essential component of the negentropic system.

Contemporary AI has a capability we underestimate: it understands context. It grasps nuance. It can read a job description written in 2019 and understand that it describes a role which has since evolved. It can analyse data from different sources and identify that they are talking about the same thing in different words.

That capacity for contextualization is precisely what traditional systems lack. An ERP does not understand. It stores. It calculates. But it does not grasp that “Head of Digital Transformation” and “Chief Digital Officer” may well denote the same role.

In LENSX, AI plays several negentropic roles.

It enriches incoming data by suggesting relevant links. It observes the organization passively and detects patterns invisible to the human eye — what we call the “Satellite”. It preserves organizational memory by structuring past insights so that they remain accessible — our “Experience Graph”.

AI does not replace human intelligence. It amplifies the organization’s capacity to maintain its order. It is the metabolism that lets the system stay negentropic.

The First Results

We have deployed this approach in several contexts: African governments in transformation, large European banks, industrial companies being restructured.

The results observed are consistent with the negentropic hypothesis.

Reconciliation time falls drastically. Leaders who used to spend hours understanding the state of their organization now reach that understanding in a few clicks. The energy released is reinvested in what matters: analysing, deciding, acting.

Organizational memory builds up. Past decisions, their contexts, their reasons, stay accessible. New teams no longer have to reinvent what their predecessors had learned.

Invisible links become visible. Dependencies between projects that nobody knew about emerge. Risks are identified before they materialize. Opportunities for synergy appear.

But the most striking result is qualitative. The leaders who use the system describe a sense of “clarity regained”. As though a fog had lifted. As though the organization had suddenly become legible to them.

That, we think, is the sensation of negentropy. The sensation of order emerging from chaos.

The Limits Of The Experiment

It would be dishonest to claim we have solved the problem of organizational entropy. Our work remains an experiment in progress.

Several questions stay open.

Is negentropy durable? Biological systems have to import energy constantly to maintain their order. What about organizational systems? Does LENSX maintain order permanently, or will entropy eventually regain the upper hand in a form we have not anticipated?

Does the approach generalize? Our deployments involve organizations of a few thousand to a few tens of thousands of people. Do negentropic principles work at the scale of a multinational with 200,000 employees? We do not know yet.

What are the second-order effects? A system that reduces entropy necessarily changes the dynamics of power in the organization. Information is power. What happens when information becomes universally accessible? We observe resistance, sometimes fierce, from those whose role consisted precisely in holding and controlling information.

These questions are not objections. They are the research programme for the years ahead.

What Physics Teaches Management

Beyond our particular experiment, applying the concept of entropy to organizations opens wider perspectives.

It suggests that organizational disorder is not a failure of management. It is a natural law. Any organization, left to itself, will tend towards chaos. Not out of incompetence, but out of physics.

It implies that point interventions — reorganizations, audits, transformations — are structurally insufficient. They inject order momentarily, but they do not change the fundamental dynamic. To counter entropy you need a continuous process, not discrete events.

It indicates that unifying information is not an IT project. It is a thermodynamic necessity. As long as information stays fragmented, entropy will keep increasing. “Data projects” are not technical initiatives. They are attempts — often clumsy ones — to create negentropy.

And finally it illuminates the potential role of artificial intelligence in organizations. Not as a tool of automation or a conversational gadget, but as a component of the negentropic metabolism. A capacity to contextualize, connect, and preserve that exceeds what humans can do.

Conclusion: The Entropic Choice

Every organization faces, consciously or not, a fundamental choice.

Accept entropy as a given. Carry on injecting order at intervals. Exhaust itself in reconciliation meetings. Watch transformations bog down. Watch complexity grow inexorably.

Or try something else. Explore the possibility of a negentropic system. A system that does not fight entropy episodically, but reduces it continuously. An organizational metabolism.

We do not claim to have found the definitive answer. LENSX is an experiment, not a certainty. But the first results suggest Schrödinger had it right. That negentropy is possible. That order can be maintained.

And that organizations, like living organisms, can choose not to die of their own complexity.