---
title: "Chapter 2: The Department Was a Workaround"
book: "Collapse or Collapse"
author: "Vinay Pasricha"
url: https://vinaypasricha.com/books/collapse-or-collapse/chapter-2-the-department-was-a-workaround
status: "On Kindle: https://www.amazon.com/dp/B0HNL53J5R. The full text is shared on vinaypasricha.com with the author's permission."
---

# Chapter 2: The Department Was a Workaround

Four men’s handwriting on one page is the whole history of management.

I did not think that when I stood in the stores with the ledger open. I thought it thirty-five years later, when I had run companies of my own, built departments of my own, and watched a machine do in an afternoon what one of those departments had done in a month. But the page had already said it. Each man had written what he knew and handed the page to the next, because no one of them could hold the whole. The whole was the reorder level of one bronze bush: what would sell, what that needed, what was on the shelf, what was on its way. Four kinds of knowledge, four heads, four hands, and a month between the question and the answer.

That is a department. Not the room, not the sign on the door, not the man who runs it. A department is what you build when a result needs more than one person can hold, and you cut the result into the pieces that single people can hold, and you put a desk under each piece. Everything else about it, the hierarchy, the reports, the meetings, the budget, the career, the business school that teaches you to run it, grows out of that one cut.

This chapter is about the cut. Why it was made. What it has cost, in money and in outcomes, for two hundred years. Why every discipline of management you were taught is a sub-discipline of it. Why it no longer needs making. And why I expect that, within five years, every company that has not undone it will be undone by one that has. I am going to say these things without hedging, because they are not opinions. They are what follows from one constraint and its removal, and the constraint has been removed.

## The constraint

Start with what a company is for. It exists to produce outcomes for someone outside it: a machine that ships, a loan that is approved, a patient who goes home, a hire who joins. Each of these is a single thing to the person who receives it. The customer does not experience sales and then production and then dispatch; he experiences a machine arriving or not arriving. The outcome is whole on the receiving end.

It is on the producing end that it comes apart, and it comes apart for a reason so old and so obvious that nobody questions it. An outcome needs many skills. The machine that ships needs someone who can forecast, someone who knows the plant, someone who keeps records, someone who finds mistakes, someone who counts shelves. A loan needs someone who can read a balance sheet, someone who knows the law, someone who knows the customer, someone who knows the bank’s appetite that month. No one person has all the skills, and even if one did, no one person has the hours. So the outcome is divided into the parts a single person can carry, and each part is given to a specialist, and the specialists are grouped by what they are good at, and the groups are called departments.

Note what has happened. The company has stopped being organised around the thing the customer feels and started being organised around the things its people are good at. That is not a small change. It is the whole change. Every management writer of the last two hundred years has been describing its consequences, and almost none of them named its cause, because the cause was invisible in the way that water is invisible to a fish. No one person could hold the whole. It was simply true, everywhere, always, and so nobody wrote it down.

I am writing it down, because it has stopped being true.

## Two centuries of the workaround

The workaround did not arrive at once. It hardened over two hundred years, and each generation of the best minds in business added a layer to it, honestly, because each of them was answering the same constraint with the best tools of their time. I will take the arc quickly, because the arc is the point: every step made the cut deeper, and every step was right.

**Smith named the gain.** In 1776 Adam Smith watched ten men in a pin factory divide the making of a pin into about eighteen operations, one man drawing the wire, another straightening it, a third cutting it, and found that they made upwards of forty-eight thousand pins a day where a man working alone, he judged, could scarcely have made twenty.[^1] Division of labour multiplied output two-hundredfold, and Smith was right, and the world took his lesson and ran with it for two centuries. What he did not dwell on, because it was not his subject, was that the pin now belonged to nobody. Ten men had made it and none of them had made a pin.

**Taylor named the task.** A century and a quarter later Frederick Winslow Taylor took Smith’s division down to the level of the single motion, timed it with a stopwatch, and gave the result to management: the worker was to do what the card said, and the thinking was to be done elsewhere, by people whose job was thinking.[^2] Scientific management was the cut made explicit and made permanent. The person who held the skill and the person who held the purpose were now, by design, different people, and a new layer was invented to stand between them. That layer is still there. It is called management.

**Ford built the cut into the floor.** The moving assembly line of 1913 took Taylor’s tasks and arranged them in a sequence so that the work came to the man rather than the man to the work, and the time to assemble a chassis fell from over twelve hours to ninety-three minutes.[^3] Ford’s line is the department made physical: every station a specialty, every handoff a few feet of chain, the whole outcome visible only from the gallery above. It is also, and this will matter later in the book, the first system in which a stall anywhere stopped everything and everyone could see it. The line did not tolerate silent failure. Offices, which copied everything else from Ford, did not copy that.

**The corporation drew the chart.** By the middle of the twentieth century the workaround had a picture of itself: the organisation chart, boxes in rows, lines between them, a function in every box and a head on every function, divisions above the functions and a headquarters above the divisions. The chart was not a description of how outcomes were produced. It was a description of how the workaround was staffed, and it became so universal that within a generation nobody could imagine a company without one. Ask a child today to draw a company and she will draw boxes. The boxes are a hundred years old and they were never the point.

**Drucker named the worker who came after the line.** The knowledge worker of 1959 carried his means of production in his head, could not be timed with a stopwatch, and had to be managed, Peter Drucker said, by objectives rather than by tasks.[^4] This was the office finding itself. The forecaster, the planner, the analyst, the clerk were knowledge workers, and the company organised them exactly as Ford had organised fitters: by what they knew, in functions, with a manager above each function to carry the purpose the function could not see. Drucker understood better than anyone that the function was a problem. He spent fifty years telling managers to look at the customer and the result rather than at their own department. He could not tell them how to remove the department, because the constraint that created it still held.

**The business school taught the workaround as a science.** Through the second half of the century the management of the cut became a profession, with its own degree, its own journals, its own consultancies, and its own certainty. I sat in its classrooms myself, at Harvard, on the programme it runs for owners of companies, and I will say in a moment what it taught me. For now it is enough to notice what the MBA is: a two-year education in how to run a company that has been cut into pieces, taught by people who never saw a company that had not been. Organisational design is the science of where to make the cut. Span of control is the science of how many pieces one manager can hold. Finance is the science of allocating money across the pieces. Human resources is the science of staffing them. Operations is the science of the handoffs between them. Not one course in the curriculum asks whether the pieces need to exist. The constraint was so complete that the question could not be formed.

**Hammer tried to undo the cut without the machine.** In 1993 Michael Hammer and James Champy published *Reengineering the Corporation*, and for a moment the whole of management stood at the edge of this book.[^5] They saw the leak at every handoff. They said, in so many words, that companies should be organised around outcomes rather than tasks, that one person should carry a whole process where possible, and that the functional organisation was a relic of Smith and Taylor. They were right about all of it. And reengineering failed, by their own estimate, in half to seventy per cent of the companies that tried it, for one reason that they did not have the tools to see: they were asking one person to hold the whole, and one person still could not. The skills still lived in heads. You could redraw the chart around the outcome, but the outcome still needed seven heads, and the seven heads re-formed into seven desks before the consultants’ invoices were paid. Reengineering was the collapse attempted thirty years early, without the machine that makes it possible. It is the strongest proof I know that the constraint, and not the chart, was always the thing. Hammer had already named the other way to fail, three years before the book, and chapter 5 takes it up: paving the cow path.

So the workaround hardened. Smith divided the outcome, Taylor separated skill from purpose, Ford laid the pieces in a line, the corporation drew the chart, Drucker tried to manage the pieces back towards the whole, the business school made the whole apparatus a science, and Hammer tried to tear it down with his bare hands and could not. Two hundred years, seven honest answers to one constraint, and at the end of it an organisation that every company in the world recognises as natural, as though companies had always come in boxes, and as though the boxes were the point.

They were never the point. They were the price.

## The price

Here is the price, in the only unit that matters, which is what the customer feels, and then in the unit the owner feels, which is money.

Every time an outcome crosses from one desk to the next, three things leak. Context leaks: the second desk knows less about the customer than the first, and the fifth knows almost nothing. Urgency leaks: the man who took the customer’s call felt it, the man who received the file did not, and by the fourth handoff the file is one of a pile. Ownership leaks: at every desk there is someone who did his part correctly, and at the end there is a machine on the floor that cannot ship and nobody who is wrong. I watched all three leak in 1989 through a chain of five men who were, every one of them, good at their jobs.

I have watched it leak since through chains of twenty, in every industry I have touched, and I will give you five, because the reorder book was a small company and you may be thinking that the problem was size. It was not. Size makes it worse, and so does sophistication.

**A hire.** A client needs an engineer, and the outcome is an engineer who joins and stays. In an ordinary recruitment firm the client’s brief goes to an account manager, who writes it down. The written brief goes to a sourcer, who has never met the client, and who searches on the words in the brief rather than the need behind them. The sourcer’s list goes to a screener, who has met neither client nor sourcer, and who filters on the words again. The screened list goes to a coordinator, who schedules interviews and knows nothing about engineering. The interviews are done by an interviewer who has read the brief once. The chosen candidate goes to an offer desk, which has never spoken to anyone involved and negotiates from a form. Seven desks. At each of them the brief is a little further from the client who gave it, a little less urgent, a little more somebody else’s. The client experiences one thing: three weeks of silence, then a candidate who is not quite what he asked for, then another three weeks. He does not know there were seven desks. He knows the outcome was not delivered, and he blames the firm, and he is right to, because the firm chose the seven desks.

**A loan.** A small business needs working capital, and the outcome is money in its account before the order it needs the money for is lost. The application goes to a relationship manager, who knows the customer. It goes to a credit analyst, who knows the numbers and not the customer. It goes to a risk committee, which knows the policy and neither. It goes to legal, which knows the documents. It goes to operations, which disburses. Six desks, six weeks, and a yes that was knowable on the first morning by anyone who held the customer, the numbers and the policy in one head. The customer, meanwhile, has lost the order, and takes the money from a lender who answered in a day, at twice the rate, and the bank has lost a customer it will never know it lost, because no desk in the chain will ever learn that the order was the point.

**A discharge.** A patient is well at nine in the morning, and the outcome is that she is home by lunch. The doctor writes the order. The nurse prepares the file. The pharmacy prepares the medicines. The billing desk prepares the account. The transport desk arranges the car. Each waits for the one before, each has twenty other patients, each does its part correctly, and the patient who was well at nine leaves at six, having occupied for nine hours a bed that another patient was waiting for in a corridor. Nobody was slow. The chain was long.

**A claim.** A car is damaged, and the outcome is that it is repaired and the owner is paid. The claim is registered by one desk, assessed by a second, surveyed by a third, approved by a fourth, and paid by a fifth, and at every desk a document is requested that the previous desk already had, because the desks do not share a head. The owner, who was promised a decision in seven days, receives one in thirty, and the insurer, which has spent more on the five desks than on the repair, wonders why its customers do not renew.

**A licence.** A citizen needs a permission from the state, and the outcome is the permission. I will not count the desks, because you have stood in that queue and you know. I will say only that the state is the purest form of the workaround, because it was designed by people who assumed the constraint was eternal, and so it cut every outcome into the maximum number of pieces that could be held by the minimum ability, and called the result accountability. It is not accountability. It is the leak made into a building.

Now the owner’s unit. Count the joints.

In the United States in 2014 there were 23.8 million managers, first-line supervisors and administrators, which is one for every 4.7 employees.[^6] Their work, honestly described, is the carrying of outcomes across desks: reviewing, approving, coordinating, reporting, reconciling what one piece did with what the next piece needs. Hamel and Zanini put the cost of this excess at more than three trillion dollars a year, about seventeen per cent of GDP, and estimated that a sixth of every non-manager’s time went on the internal compliance the managers required.[^7] Microsoft, reading the telemetry of its own tools, found in 2023 that across the Microsoft 365 applications it measures, fifty-seven per cent of recorded activity time went to meetings, chat and email, and forty-three per cent to creating anything.[^8] Not all communication is waste, but in the tools that record it, more than half of the time is the sound of the joints. And Gallup, which has measured it for a quarter of a century, finds that about one worker in five, worldwide, is engaged in their work at all, and that the two-point fall in that figure in 2024 alone cost the world economy, by its estimate, four hundred and thirty-eight billion dollars.[^9] That is intention leaking, measured, and the pattern repeats in every region Gallup measures, because the structure that leaks it is the same.

Put the three together and the twentieth-century company reveals itself: an organisation in which one person in five manages, in which more than half of everyone’s time is spent on the joints between desks, and in which four people in five have stopped wanting the outcome. It was a machine for producing outcomes that was mostly made of the joints between its parts, and most of what it spent was spent on the joints.

I want to be fair to the workaround, once, and then I will not be fair to it again. It built the modern world. There was no other way to make a pin, a car or a loan at scale with the people and tools available, and the men who designed it were not fools; they were the best minds of their time answering a real constraint honestly. Dhawan carrying the plant in his head was not a failure of imagination. It was the limit of a head. The department deserves the respect we give to any tool that did its job for two centuries. What it does not deserve is to be mistaken for the job.

## What the MBA taught, and why it is now wrong

I said I would say what the business school taught me. Here it is, discipline by discipline, with what each one is when the constraint is removed. I am not attacking the schools. I am describing what they will have to become, because every one of these disciplines was a science of the cut, and the cut is gone.

**Organisational design** taught where to make the cut: by function, by product, by geography, by matrix, and the trade-offs between them. It is a rich and subtle subject, and every word of it assumes that the outcome must be divided. When one system holds every skill, there is one design, and it is the outcome itself. Organisational design becomes the design of chains, which is the tuner’s craft, and the design of playbooks, which is the master’s. The question of where to cut has no answer because there is nothing to cut.

**Span of control** taught that a manager can hold five to nine direct reports, and that the pyramid follows from that number: layers equal to the logarithm of headcount. The number was true. It was the limit of a human’s attention over other humans doing pieces of work she could not see. When the work is visible as a floor of tiles and the only things that reach the manager are the things that need her, the span is a thousand, and the pyramid is two layers tall, and it will not grow a third because there is nothing for a third layer to carry.

**Functional strategy** taught marketing strategy, operations strategy, HR strategy, IT strategy, each with its own professors, each optimising its own piece. There are no functions. There are Outcomes, and each has one strategy, which is the result it exists to deliver and the budget it has to deliver it. The work that functions used to do divides cleanly in two. Most of it, a compliance check, a credit check, a quality check, the screening of a hire, is a skill used in producing someone else’s result, and a skill belongs inside the chain as an agent, not in a department beside it. A little of it, raising capital from the market, say, produces a result in its own right, and that becomes an Outcome like any other, with its own vectors and its own master. Functional strategy becomes one question asked case by case: is this a result, or a skill? The answer decides whether it is an Outcome or an agent, and no answer gives it a department.

**Human resources** taught recruitment, performance management, compensation and development as four systems for keeping a large number of people of ordinary intention pointed roughly at a result they could not see. The collapsed company needs very few people, selected for extraordinary intention, very highly paid, whose attention is read continuously from the work they do and never from watching them. Recruitment becomes selection for intention, which is a different search. Performance management becomes the daily review, which the aligner holds herself. Compensation becomes pay that follows the vector. Development becomes deeper, not higher, because there is no higher. Of the four systems, one survives in a new form and three dissolve.

**Operations management** taught the handoff: process mapping, queueing, bottlenecks, lean, six sigma, the whole discipline of making the pieces flow between desks with less waste. It was the most honest of the disciplines, because it alone admitted that the joints were where the cost was, and it spent fifty years making the joints a little cheaper. Operations management becomes the chain, which has no handoffs because it has no desks, and the gates, which are the only places the flow stops, and the stall alarm, which Ford had and the office never did. What remains of the discipline is the automation line: the question of which step a machine can hold to standard today. That question moves every year, and it is the tuner’s.

**Culture and leadership** taught that an organisation is held together by shared values, that a leader’s job is to inspire, and that engagement is the measure of both. Culture was a substitute for visibility. When a manager could not see the work, he needed people to share his values so that they would do the right thing unseen. In the collapsed company every vector is visible in a cockpit, every cockpit in a control room, every Outcome in the command centre, and nothing is done unseen, so nothing needs to be held together by shared feeling. What holds it is alignment, pay and the gates. Leadership becomes keeping, which is a quieter and rarer craft, and inspiration is replaced by selection: you do not inspire intention into people, you find the people who have it and pay them for it.

**Performance measurement** taught the balanced scorecard, the KPI, the monthly review, the quarterly report. All of it was a way of reconstructing, after the fact and at a distance, what the work was doing, because nobody could see the work. When the work is one live truth as of the last minute, flowing from cockpit to control room to command centre with nothing summarised by a person on the way, there is no report, because there is nothing to reconstruct. The board sees what the Keeper sees, at the same moment, and the entire apparatus of reporting goes, with the departments that wrote it.

Seven disciplines, and every one of them a science of managing the cut. I am not saying the schools taught nothing. I am saying they taught, brilliantly, the management of a constraint, and the constraint has been removed, and a management education now has to begin where this chapter begins: with the fact that no one person could hold the whole, and the fact that something now can.

## The constraint is gone

Now the thing that changes everything, stated as plainly as I can state it.

The constraint that created the department was that no single person could hold every skill an outcome needed. In 1989 I removed a small piece of that constraint with a program that held what four men had held between them, and four desks went quiet. Nobody was let go; all four found better work in the same building. The program could hold arithmetic. It could not hold judgement, or words, or a conversation with a customer, so the chairman kept his week and the firm kept its sales office and its purchase office and everything else that needed a human skill. The collapse went as far as the machine could hold, and stopped.

The machine can now hold almost all of it. A chain of AI agents holds the skills of a researcher, a writer, a caller, a seller, a negotiator, a coordinator and a clerk. It reads a brief and searches on the need rather than the words. It screens, schedules, interviews, drafts an offer and chases a signature, and it does each of these to the standard of a competent professional, around the clock, in the customer’s time zone, without the brief degrading between steps, because there are no steps between which it can degrade. The seven desks of the recruitment firm are one system. The six desks of the loan are one system. The five desks of the claim are one system. Where a skill is not yet held to standard, and there are such skills, the line sits a little lower this year and moves up next year, and it has never once moved down.

The capability is not a forecast. It is being measured. When seven hundred and fifty-eight consultants at one of the world’s most expensive firms were given an AI assistant for ordinary consulting tasks, those with the assistant completed twelve per cent more tasks, twenty-five per cent faster, at forty per cent higher quality than those without, and the consultants who had been weakest gained the most.[^10] The same experiment carried a warning the doctrine takes seriously: on one task chosen to lie just outside what the model could do, the consultants who used it were less likely to reach the right answer than those who did not. The line moves, and it is not yet everywhere; a gate exists for the side of it the machine has not reached. That was 2023, with a tool that was a toy beside what exists now, and with the agent assisting a human at a desk rather than replacing the desk. When a payments company put an AI assistant in front of its customers in 2024, it handled two-thirds of all customer-service conversations in its first month, two million three hundred thousand of them, in thirty-five languages, doing the work of seven hundred full-time agents, with resolution times cut from eleven minutes to two and repeat enquiries down by a quarter, at a customer-satisfaction score equal to the humans it replaced.[^11] Seven hundred desks became one system in a month. That is a collapse, measured, and it is two years old.

I will argue the claim properly later in the book, against every objection I have heard, and I will answer the loudest of them before this chapter ends. For now the consequence follows from the statement, and the consequence is this.

If the skills sit in one system, there is no reason to cut the outcome into pieces. If there is no reason to cut it, there is no reason for the desks that held the pieces. If there are no desks, there is nothing for the layers above them to carry across. The department, the function, the handoff and the manager were all built on one constraint, and the constraint has gone, and they are standing on nothing. They will stand for a while longer, because buildings stand for a while after their foundations go. Then they will not.

What is left when they go is small and precise. An outcome, whole, the way the customer feels it. A system of agents that carries it from first intention to final result. And one human, whose job is not any of the skills, because the system has those, but the one thing the system does not supply: intention. Someone who wants this outcome, for this customer, today, and will stand at the gates and say go, and will sign. The whole company collapses towards that person, and the next chapter is about her.

## The excuses

I have put this argument to owners for three years, and I have heard every reason for not acting on it. There are seven. None of them survives, and I will take them in the order they are usually offered.

**“My industry is different.”** It is not. Every industry has a connected part and a hands part. The connected part is everything that reaches a screen: every brief, application, claim, order, record, payment, message and signature. The hands part is the surgery, the shop floor, the site. The collapse applies to the connected part completely, today, in every industry without exception, and the hands part shrinks every year as the hands connect. If your industry is different, it is different only in the size of its hands part, and the hands part is not where your departments are. Your departments are in the connected part. They are the connected part.

**“Our people are our greatest asset.”** Then pay them like it. The collapsed company pays its people very highly, far above the market, because it needs a tenth as many, and the ones it keeps are the ones whose intention held. The old company said people were its greatest asset and paid them like a cost, because it needed so many of them that it could afford nothing else. If you mean the sentence, the collapse is how you honour it. If you do not mean it, the collapse is how you stop pretending.

**“AI makes mistakes.”** So did Verma, fifty times a month, and so did the register, and so did every desk in every chain I have described, and the mistakes were invisible because nobody could see across the desks to find them. The chain’s mistakes are visible, counted, logged and replayable, and they are caught at gates by a human whose only job is to catch them, and the chain does not make the same mistake twice because the tuner fixes it that day. The question is not whether the machine makes mistakes. It is whether you can see them. You could never see the desks’. You can see the chain’s.

**“We are regulated.”** Then you have an advantage, because a regulator wants exactly what the chain provides: every decision logged, every authority limit explicit, every commitment signed by a named human under a stated permission, and an audit trail that can be replayed to the agent that produced any output. The old company gave its regulator a filing cabinet and a promise. The collapsed company gives it the engine. The signature rule in this book exists because of regulation, not despite it.

**“We tried reengineering and it failed.”** You tried the collapse without the machine, and it failed because the seven heads re-formed into seven desks, as I have said. The thing that was missing in 1993 is the thing that is present now. If reengineering taught you anything, it should have taught you that the chart was never the problem, and that the moment the constraint went, the thing Hammer wanted would work. The constraint has gone.

**“Our customers want to deal with humans.”** Your customers want the outcome. They want the engineer who joins, the money in the account, the car repaired, the permission granted. They never wanted the seven desks; they tolerated them because every supplier had them. When one supplier does not, your customers will not ask whether a human read their brief at desk four. They will ask why yours took three weeks. And when something must be signed, a human will sign it, and the customer will know her name.

**“We are too small,” or “we are too big.”** The thirty-person plant collapses its office into one vector and keeps its floor until the floor connects; the ten-thousand-person company collapses into a dozen Outcomes of a thousand vectors each, held by a dozen masters, under one Keeper. Size changes the number of vectors. It changes nothing else. The reorder book was a small company. Klarna, in chapter 19, was not.

There is an eighth reason, which nobody offers aloud and everybody means, and I will offer it for them. “I built the departments. The departments are my career, my status, my understanding of what a company is. If they go, what am I?” You are the Keeper, if your intention holds. If it does not, you are what the department was: a workaround, and the next chapter is not about you.

## What happens to the people

You will want to know what happens to the people, and I will not make you wait for chapter 13 to hear the shape of it. There will be far fewer of them, and they will be paid far more, because what you are buying from them has changed from skill, which is now cheap, to intention, which has always been rare. The rest will be released. In 1989 the collapse gave time back, because the manual frontier was vast and four freed men walked across it, inside the same company, to new desks and the next unautomated job; nobody was let go. There is less of that frontier now and there will be less every year, and a book that told you otherwise would not be worth your reading. What the doctrine owes those people, and what the companies that collapse owe them, I take up where it belongs, in chapter 13, in the month the headcount falls, and again where an Outcome dies. For now, only this: the alternative to collapsing deliberately, around the people whose intention holds, is being collapsed carelessly, around whoever is left.

## The company down the road

Because here is what the owner reading this has to understand, and it is not a prediction. It is a description.

Somewhere near you, perhaps already, there is a company that does what yours does and has never built a department. I am going to show you a day in it, and then its customer’s day, and then its books, because you will not believe it until you have seen all three.

It has five people. One of them, call her the aligner of Winning, holds the outcome of winning customers in your market. One holds the outcome of delivering what was won. The other three are the same two things for a second market and a tuner who stands by for all four, which the doctrine itself counts as a gap: chapter 5 gives every vector a tuner of its own. Each has a chain of agents: for Winning, an agent that finds prospects, one that qualifies them, one that writes the first approach, one that handles the replies, one that drafts the proposal, one that negotiates within a limit she set, and one that chases the signature. The chain works all night in your customers’ time zones. At six in the morning she puts on a headset, and an agent that answers for the whole chain tells her what happened: five hundred prospects found, four hundred and sixty qualified and waiting at a gate, thirty replies drafted and waiting at another, one proposal accepted with a change in terms and held for her signature. She samples, releases, holds one, signs one. Forty minutes. She does none of the steps. At noon and at six she does twenty minutes more. The rest of the day is hers. There is no manager above her, because there is nothing between her and the master to manage. There is no function beside her, because there are no functions. There is no meeting, because there is nobody to meet.

Now her customer’s day, which is your customer’s day, because they are the same person. On Monday morning he describes what he needs, once, to one chain, in his own words. By Monday afternoon the chain has understood the need behind the words, because it searched on the need and not the words, and has sent him three candidates, or three terms, or three dates, whichever his outcome requires. On Tuesday he replies. His reply does not go to a desk; it goes to the chain, which has the whole of his brief and the whole of his history in one place, and the chain answers within the hour, in his time zone, and the answer is as new as that hour. On Wednesday he asks for a change. A human reads it, decides, and signs, and he knows her name, and she is the same human who will answer him next month, because the outcome is hers. He does not know, and will not care, that there were no desks. He knows that the thing he asked for arrived in three days with no leak, and that the last time he asked your company it arrived in three weeks with several. He is not going to come back to you. Not out of anger. Out of arithmetic.

Now the books. Your company employs forty-five people to produce what those five produce, because that is what seven desks per outcome, two markets, two outcomes and the managers to carry across them come to. Your forty-five are paid at the market, with a building, with the sixteen per cent of their time that goes on the compliance the managers require, and with one in five of them managing. The five down the road are very highly paid, and still their whole payroll is a fraction of yours; they have no building, no managers, no reporting, and a chain whose running cost is a fraction of one salary. They are cheaper than you before a single outcome is counted. Then count the outcomes. Their customer gets a brief that does not leak, an answer that is never a month old, and a human who signed it, and he stays. Yours leaves. Their revenue per person is nine times yours before the leak is counted and more after. Their margin is not better than yours. It is a different kind of number.

That company is not better than yours. Its people are not cleverer than yours. It simply never had the constraint, and so it never built the workaround, and so it has none of the joints that you are paying for. Everything you spend on the joints, it spends on the outcome. That is the entire difference, and it is enough.

## Five years

I will put a date on it, because a doctrine that will not be held to a date is not a doctrine.

I expect that within five years every connected outcome in every industry will be run this way by someone. Not by everyone; by someone, in your market, selling to your customers, at a tenth of your cost, with none of your leaks. In some industries it is already so, and you have not noticed, because the company doing it is small and quiet and does not look like a competitor; it looks like five people. It is the most dangerous competitor you have ever had, because you cannot match it by trying harder. You cannot manage your way to it. You cannot reengineer your way to it; that was tried. You can only collapse to it, and the collapse takes a year, and the year has to start before the five people reach your customer.

Everything I have told you in this chapter follows from one fact and its negation. For two hundred years, no one person could hold the whole, and so we cut it, and the cut became the company, and the company became a science, and the science became a profession, and the profession became your career. Now something can hold the whole. Not a person; a system, with a person at the gate. The cut has no reason left. The departments are standing on nothing, and the people who built them, who are honest people answering an honest constraint, are about to learn what Mr Malhotra learned when he picked up my drawing of his own projection sheet: that the thing they do with their best week every month is not, after all, the thing that only they can do.

You can be that company. The departments you have are not your company; they are the scaffolding that was put up because the company could not stand without it, and the company can stand without it now. Take the scaffolding down yourself, in the order the method describes, around the people whose intention you would pay very highly for, and run old beside new until the registers go quiet. Or leave it up, and wait for the company down the road to do it for you.

Collapse your organisation now, or your organisation will collapse.

[^1]: Adam Smith, *An Inquiry into the Nature and Causes of the Wealth of Nations* (1776), Book I, Chapter 1. Smith’s figures: ten men, about eighteen distinct operations, upwards of forty-eight thousand pins a day; a man working alone “could scarce, perhaps, with his utmost industry, make one pin in a day, and certainly could not make twenty.”

[^2]: Frederick Winslow Taylor, *The Principles of Scientific Management* (1911).

[^3]: Henry Ford’s Highland Park plant introduced the moving chassis assembly line in late 1913; Ford’s own account (*My Life and Work*, 1922, with Samuel Crowther) gives chassis assembly time falling from twelve hours and twenty-eight minutes to one hour and thirty-three minutes.

[^4]: Peter F. Drucker introduced “management by objectives” in *The Practice of Management* (1954) and the term “knowledge worker” in *Landmarks of Tomorrow* (1959).

[^5]: Michael Hammer and James Champy, *Reengineering the Corporation: A Manifesto for Business Revolution* (1993). The authors’ own estimate, given in the book, was that fifty to seventy per cent of reengineering efforts failed to achieve the dramatic results intended; later studies put the failure rate at seventy per cent or higher.

[^6]: Gary Hamel and Michele Zanini, “Excess Management Is Costing the U.S. $3 Trillion Per Year,” *Harvard Business Review*, September 2016: 23.8 million managers, first-line supervisors and administrators in the US workforce in 2014, one for every 4.7 employees.

[^7]: Hamel and Zanini, as above: excess bureaucracy costing more than $3 trillion a year, about 17 per cent of GDP; roughly 16 per cent of non-managerial employees’ time spent on internal compliance.

[^8]: Microsoft, *Work Trend Index Annual Report: Will AI Fix Work?*, May 2023: across Microsoft 365 users, 57 per cent of measured time went on communicating (meetings, email and chat) and 43 per cent on creating (documents, spreadsheets and presentations). The figure covers activity in Microsoft’s tools, not every hour of the working week.

[^9]: Gallup, *State of the Global Workplace* reports, 2024 and 2025 editions: global employee engagement 23 per cent in 2023 and 21 per cent in 2024; manager engagement 27 per cent in 2024; the two-point fall in global engagement in 2024 estimated to have cost US$438 billion in lost productivity (Gallup estimates that a fully engaged workforce would add US$9.6 trillion a year).

[^10]: Fabrizio Dell’Acqua, Edward McFowland III, Ethan Mollick, Hila Lifshitz-Assaf, Katherine Kellogg, Saran Rajendran, Lisa Krayer, François Candelon and Karim Lakhani, “Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality,” Harvard Business School Working Paper 24-013, September 2023, published in *Organization Science* 37, no. 2 (2026): 403–423, doi 10.1287/orsc.2025.21838: 758 BCG consultants; those using GPT-4 completed 12.2 per cent more tasks, 25.1 per cent faster, at over 40 per cent higher quality; the lowest-performing half improved by 43 per cent. On the task designed to fall outside the frontier, consultants using AI were 19 percentage points less likely to produce a correct solution than those working without it.

[^11]: Klarna, press release of 27 February 2024, “Klarna AI assistant handles two-thirds of customer service chats in its first month”: 2.3 million conversations in the first month, two-thirds of all customer service chats, across 23 markets and 35 languages, equivalent to the work of 700 full-time agents; average resolution time under two minutes against eleven previously; a 25 per cent drop in repeat enquiries; customer satisfaction on par with human agents; an estimated US$40 million profit improvement in 2024.