In 2018 I wrote an article arguing that organizations sit somewhere on a continuum between two failure modes: too much entropy and too little. I called the alternative the Intentional Organization. I didn’t realize it at the time, but I was writing the founding document of the company I’d start years later. So consider this less a revision than a return — to an idea I now understand better than I did when I first had it.
Here’s the idea, sharpened by several more years of watching organizations succeed and fail at it: an organization is not a structure that occasionally changes. It is a continuous fight against disorder that occasionally produces structure. Get that backwards — as almost everyone does — and no amount of reorganizing, strategy-setting, or transformation programming will save you.
The physics, taken more seriously this time
Entropy is the tendency of every system to drift from order toward disorder unless energy is continuously supplied. The teenager’s clean room reverts to chaos. The campfire becomes ash and heat. The universe, given enough time, smooths out into undifferentiated nothing.
Organizations are not exempt. Left untended, alignment decays. Communication thins. Shared purpose fades into a poster in the lobby. Two teams develop incompatible understandings of the same strategy. None of this requires anyone to do anything wrong — it requires only that no one continuously do the work of holding shared clarity together. Disorder is the default. Order is the achievement.
In the original article I framed the response to this as a continuum with two bad endpoints: the Entropic Organization (all frenetic motion, no coherence) and what I called the Negentropic Organization (frozen solid, all structure, deathly afraid to melt). That framing still mostly works, and I’ll return to those failure modes in a moment. But the physics I was reaching for has a better model, and it changes the prescription.
Living systems don’t survive by minimizing entropy. They survive by metabolizing it. An organism is not a sealed, stable structure — it’s an open system that maintains its order by continuously taking in energy and information, processing it, and expelling disorder back into the environment. Stop the flow and you don’t get stability. You get death. Schrödinger made a version of this observation about life itself; complexity scientists later showed that all sorts of ordered systems — whirlpools, weather cells, cities — exist only because energy is flowing through them, not despite it.
This is the upgrade the 2018 version of me needed: an organization is not a building that erodes and needs maintenance. It’s a candle flame — a pattern that persists only as long as something is flowing through it. The question is never “how do we stop the flux?” The question is “do we have the capacity to metabolize it?”
The two refusals
Seen this way, the two failure modes from the original article turn out to be the same failure wearing different costumes. Both are refusals to metabolize.
The dissolving organization refuses by drowning. There is enormous motion — initiatives launched, reorgs announced, Slack channels multiplying, everyone heroically busy — but no mechanism for digesting any of it. Information pours in and passes straight through, unprocessed. Lessons aren’t extracted from failures because no one stays with anything long enough to extract them. Busyness becomes the badge of honor; burnout, the membership fee. From the outside it can look like vitality. It’s actually the organizational equivalent of a fever: lots of heat, no work being done.
The frozen organization refuses by sealing itself off. Leaders sense the chaos and respond by locking everything down — policies for every contingency, approval chains for every decision, a yearly plan defended against all evidence. The organization achieves a kind of order, and that order feels like safety. It is safety, in exactly the way an ice cube is safe: stable right up until the environment changes, at which point it doesn’t bend, it disappears. A new competitor, a departed leader, a supply shock, a technology shift — the frozen organization meets novelty with a structure that has no capacity to process it, because processing was the very thing the structure was built to prevent.
Most organizations aren’t purely one or the other. They oscillate — a few years of dissolution provokes a freeze; the freeze becomes unbearable and provokes a “transformation” that reintroduces chaos. The pendulum swings between two ways of avoiding the actual work.
The actual work
So what does metabolizing look like? It’s a cycle with three movements, each of which is a learnable organizational capability rather than a personality trait of lucky companies.
Noticing. All change begins with awareness that something is off — or could be better — and most organizations are structurally bad at it. The dissolving org moves too fast to notice anything; the frozen org has decided in advance that there’s nothing to notice. Noticing requires deliberately protected attention: an operating rhythm with reflection built into it (so retrospection doesn’t depend on heroic willpower), genuine transparency (so useful information can reach the people who could act on it), and — hardest of all — the willingness to stop doing things and observe what happens. Attention is the scarcest resource in any organization, scarcer than money or talent or ideas, and noticing is the first claim on it.
Sensemaking. Noticing produces raw signal; sensemaking turns it into meaning. This is fundamentally a conversational capability — the capacity for people to ask why out loud, to reason from first principles rather than from organizational habit, to say “I think what we’re seeing means X” and be argued with rather than punished. It requires enough psychological safety that curiosity outcompetes fear, and enough shared purpose that there’s a lens to interpret signals through. Sensemaking is also where organizations confront the uncomfortable truth that different people, looking at the same signal, will construct different meanings — and that this multiplicity is a feature to be worked with, not noise to be eliminated by whoever ranks highest.
Experimenting. Meaning is only worth anything if it can change how you work. The third capability is making change cheap: small, reversible, time-boxed experiments instead of grand transformations; permission structures that let teams try things without petitioning three layers of hierarchy; a cultural reflex of “is it safe to try?” rather than “is it certain to work?” Organizations that experiment well don’t have better ideas than anyone else. They have a lower cost of finding out which of their ideas are good — and the humility to let reality vote.
Then the cycle turns. Every experiment changes the organization, which changes what it’s capable of noticing, which feeds the next round of sensemaking. Expertise compounds: a team that has been running this loop for two years perceives signals that a new team doesn’t yet know are perceivable — the way a conductor hears things in a symphony that the rest of us literally cannot.
The AI complication
The 2018 version of this argument didn’t have to contend with what’s now obvious: we’ve connected every organization to a machine that produces output at near-zero cost.
This is routinely described as a productivity revolution, and it is — but notice what kind. AI is an extraordinary accelerant of motion: more documents, more analyses, more options, more plausible-sounding plans, more of everything, faster. It supplies the inputs to the metabolic cycle at a rate no organization has ever experienced. What it does not supply — cannot supply — is the metabolizing itself. Noticing what matters in the flood. Making shared meaning of it. Deciding, together, what to try.
Which means AI doesn’t relieve the entropic pressure on organizations. It multiplies it. The dissolving organization now dissolves faster, drowning in machine-generated motion it has even less capacity to digest. The frozen organization freezes harder, locking down the new tools with policies before anyone learns anything from them. The constraint on organizational performance is migrating, decisively, from producing output to metabolizing it — and the metabolic capabilities are made of exactly the things AI doesn’t provide: attention, judgment, trust, and conversation. The organizations that thrive in the next decade won’t be the ones with the most AI. They’ll be the ones with the digestive system to match their new appetite.
The Deliberate Organization
I called this the Intentional Organization in 2018. I’d name it differently now — partly because my thinking has a new home, but mostly because deliberate says something intentional doesn’t. Intention is a state of mind. Deliberateness is a practice: the ongoing, unglamorous, repeated act of choosing — where attention goes, what gets noticed, what meaning gets made, what gets tried next. You can hold an intention while doing nothing. You cannot be deliberate while doing nothing.
The Deliberate Organization, then, isn’t defined by any particular structure, model, or method. There is no perfect org chart, no perfect compensation scheme, no perfect way of working — only ways that are more or less fit for a specific organization at a specific moment, all of them temporary. What defines the Deliberate Organization is that it has built the cycle — notice, make sense, experiment — into its basic way of operating, so that it can keep re-fitting itself as the moment changes. It is not a destination. It’s a direction, walked daily.
And here is the part I believe more strongly now than I did then: this capability cannot be installed from outside. A consultant can hand you a structure; nobody can hand you a metabolism. It has to be grown, by the people inside the organization, through practice — which is slower and less satisfying than buying a transformation, and also the only version that survives contact with the next disruption. The work is the work. It starts when you decide that fighting entropy on purpose beats losing to it by default.
A note on lineage: an earlier version of this argument appeared in 2018 as “Designing the Intentional Organization," written during my decade at The Ready. The entropy frame and the notice–sensemake–experiment cycle began there; the dissipative-systems reframe, the AI argument, and the shift from intention to deliberateness are where the thinking has gone since.