Originally published on LinkedIn on August 6, 2026.
"There is nothing so useless as doing efficiently that which should not be done at all."
— Peter Drucker
Something unusual has been happening inside many of the world's largest organizations.
Over the past year, companies that enthusiastically encouraged employees to use artificial intelligence have begun quietly changing course. AI budgets exhausted months ahead of schedule. Frontier models once promoted for nearly every task now reserved for specific use cases. Adoption celebrated; returns elusive. The conversation has shifted quickly to costs — tokens, inference, compute, infrastructure.1–6
These are real concerns. But I suspect they are obscuring something more fundamental — a question that organizations have been asking, without fully recognizing it, for much longer than artificial intelligence has existed.
Why do organizations repeatedly find themselves disappointed by technologies that unquestionably make individuals more capable?
This is not an AI question. It is a management question. And the answer may be hiding in a pattern that every technological revolution has produced, and that every generation of management has struggled to see clearly.
Every revolution has made the same promise
The Industrial Revolution transformed production beyond anything previously imaginable. Machines allowed organizations to manufacture at scales that would have been inconceivable only decades earlier. Entire industries were rebuilt around mechanical capability, and productivity soared in ways that permanently altered economic life.
Yet the organizations that prospered most were not simply those with the most machinery. They were those that learned how to organize production itself — how to coordinate human work, structure responsibility, manage supply chains, and develop the collective routines that turned mechanical capability into organizational output. The machine was an instrument. The organization was always the object.
A century later, the Information Age offered another extraordinary promise. Information would become abundant, enterprise software would connect functions, and knowledge would finally flow where decisions were being made. Organizations invested accordingly, and information certainly became abundant. Better decisions did not automatically follow. Many organizations became exceptionally good at collecting and storing information while remaining remarkably inconsistent at turning it into judgment. The technology delivered precisely what it promised. The organizational capacity to use what it delivered was another matter entirely.
We are now in the age of artificial intelligence. The promise sounds familiar: if every individual becomes dramatically more capable, surely the organization becomes dramatically more intelligent. The evidence emerging from enterprises around the world suggests that conclusion deserves scrutiny.
Perhaps we have been making the same mistake for generations — mistaking the instrument for the object, and wondering why the object never quite improves the way the instrument does.
What management has always been able to see
There is an understandable reason this pattern persists.
Technology is visible. Machines can be examined, purchased, implemented, and compared. Software can be deployed, measured, and presented in board meetings. AI models can be benchmarked with increasing precision. The capabilities technology introduces are immediately observable. The conditions that determine whether those capabilities improve the organization are not.
No dashboard captures the quality of collective judgment. No quarterly report measures how effectively knowledge crosses functional boundaries. No balance sheet records whether an organization is becoming better at thinking together or simply better at moving faster in the same direction.
Herbert Simon, who spent a career mapping how organizations actually make decisions, observed that organizations do not make decisions.8 People inside organizations make decisions, shaped by the structures, routines, and shared understanding that surround them. Improve the technology without improving those structures, and you improve the speed and scale at which existing patterns reproduce themselves — for better or worse.
This creates a subtle but consequential bias in management itself. Investment becomes easier to justify than integration. New capabilities become easier to celebrate than better judgment. Over time, organizations become increasingly sophisticated at acquiring technologies while remaining comparatively underdeveloped in understanding how those technologies change the cognitive demands placed on the organization. Technology has never been management's final object. The confusion between the instrument and the object may be the most persistent and expensive mistake in the history of the discipline.
Capability and complexity compound together
Every major technological revolution has produced the same surprise.
As organizations become more capable, they also become more complex. Machines increased production, but they also increased coordination requirements. Information systems increased access to knowledge, but they also increased the volume of decisions requiring interpretation.10 Artificial intelligence dramatically expands individual capability while simultaneously raising the cost — and the difficulty — of organizational coherence.
Technology does not simplify organizational complexity. It amplifies it.
This creates a paradox that sits beneath every wave of enterprise technology disappointment. The more powerful the capabilities an organization deploys, the more it depends on organizational intelligence it may never have developed. The gap between what individuals can now do and what the organization can coherently decide to do with those capabilities grows wider, not narrower, with each technological generation.
It may explain why organizations routinely report individual productivity gains that do not translate into organizational performance gains. The individuals are genuinely more capable. But capability distributed across thousands of specialized people, without the architecture to integrate it, produces something other than intelligence. It produces activity — faster, louder, and more expensive than before.
The object management missed
The two previous articles in this series each performed a specific reframe. The organization became a cognitive system, not merely an economic one. Specialization became the structural reason collective intelligence is necessary, not simply a way to improve efficiency. This article draws the implication that neither of those moves made explicit.
If organizations are cognitive systems, then the true object of management has always been the organization's capacity to think — to perceive, integrate, learn, and decide collectively. Technology has never been that object. It has always been an instrument through which that capacity gets tested, revealed, and either developed or neglected.
Every technological revolution has been, at its core, an invitation to ask a question that management mostly declined to answer. Not: how do we deploy this technology? But: what does this technology reveal about how our organization actually thinks?
A more consequential question
The current AI reckoning — budgets exhausted, returns elusive, governance scrambled — may be the clearest invitation yet to ask that question seriously.
If the limiting factor has never been the technology, and if technology has always been an amplifier of whatever organizational capacity already exists, then the question every leadership team should be asking is not which AI to deploy. It is what kind of organization they are building to use it.
That question leads somewhere most management frameworks have not yet gone — not to the technology, not to the hierarchy, not even to the talent, but to the largely invisible substrate through which people in organizations develop shared understanding, coordinate action, and become capable of thinking together.
Every organization already possesses this substrate. It exists whether leadership recognizes it or not. It shapes every meeting, every decision, every disagreement, and every strategic initiative. Yet because it rarely appears on an organizational chart or a financial statement, it remains largely invisible until moments of extraordinary change expose it.
What that substrate actually consists of — and why the most consequential decisions an organization makes may happen long before anyone enters a meeting room — is where this investigation goes next.
───
Selected References
Enterprise Evidence
1. TechCrunch. The Token Bill Comes Due: Inside the Industry Scramble to Manage AI's Runaway Costs (2026).
2. TechCrunch. Companies Are Scrambling to Stop Employees from Maxing Out AI Budgets with Small Tasks (2026).
3. Gartner. Autonomous Business and AI Layoffs Create Budget Room but Do Not Deliver Returns (2026).
4. Financial Times / Australian Financial Review. Enterprise AI governance and AI operating costs (2026).
5. Business Insider. Enterprise AI adoption, organizational incentives, and AI governance (2026).
6. Oplexa. The AI Inference Cost Crisis (2026).
Foundations of Organizational Thought
7. Hayek, F. A. (1945). The Use of Knowledge in Society.
8. Simon, H. A. (1947). Administrative Behavior.
9. March, J. G., & Simon, H. A. (1958). Organizations.
10. Galbraith, J. R. (1973). Designing Complex Organizations.
11. Weick, K. E. (1995). Sensemaking in Organizations.
12. Coase, R. H. (1937). The Nature of the Firm.
13. Williamson, O. E. (1985). The Economic Institutions of Capitalism.
14. Hackman, R. (2002). Leading Teams.
Explore the OCI series
Previous: Why Organizations Need Intelligence
Next: The Illusion of Alignment
Discuss organizational co-intelligence in your organization.
Originally published on LinkedIn on August 6, 2026.
"There is nothing so useless as doing efficiently that which should not be done at all."
— Peter Drucker
Something unusual has been happening inside many of the world's largest organizations.
Over the past year, companies that enthusiastically encouraged employees to use artificial intelligence have begun quietly changing course. AI budgets exhausted months ahead of schedule. Frontier models once promoted for nearly every task now reserved for specific use cases. Adoption celebrated; returns elusive. The conversation has shifted quickly to costs — tokens, inference, compute, infrastructure.1–6
These are real concerns. But I suspect they are obscuring something more fundamental — a question that organizations have been asking, without fully recognizing it, for much longer than artificial intelligence has existed.
Why do organizations repeatedly find themselves disappointed by technologies that unquestionably make individuals more capable?
This is not an AI question. It is a management question. And the answer may be hiding in a pattern that every technological revolution has produced, and that every generation of management has struggled to see clearly.
Every revolution has made the same promise
The Industrial Revolution transformed production beyond anything previously imaginable. Machines allowed organizations to manufacture at scales that would have been inconceivable only decades earlier. Entire industries were rebuilt around mechanical capability, and productivity soared in ways that permanently altered economic life.
Yet the organizations that prospered most were not simply those with the most machinery. They were those that learned how to organize production itself — how to coordinate human work, structure responsibility, manage supply chains, and develop the collective routines that turned mechanical capability into organizational output. The machine was an instrument. The organization was always the object.
A century later, the Information Age offered another extraordinary promise. Information would become abundant, enterprise software would connect functions, and knowledge would finally flow where decisions were being made. Organizations invested accordingly, and information certainly became abundant. Better decisions did not automatically follow. Many organizations became exceptionally good at collecting and storing information while remaining remarkably inconsistent at turning it into judgment. The technology delivered precisely what it promised. The organizational capacity to use what it delivered was another matter entirely.
We are now in the age of artificial intelligence. The promise sounds familiar: if every individual becomes dramatically more capable, surely the organization becomes dramatically more intelligent. The evidence emerging from enterprises around the world suggests that conclusion deserves scrutiny.
Perhaps we have been making the same mistake for generations — mistaking the instrument for the object, and wondering why the object never quite improves the way the instrument does.
What management has always been able to see
There is an understandable reason this pattern persists.
Technology is visible. Machines can be examined, purchased, implemented, and compared. Software can be deployed, measured, and presented in board meetings. AI models can be benchmarked with increasing precision. The capabilities technology introduces are immediately observable. The conditions that determine whether those capabilities improve the organization are not.
No dashboard captures the quality of collective judgment. No quarterly report measures how effectively knowledge crosses functional boundaries. No balance sheet records whether an organization is becoming better at thinking together or simply better at moving faster in the same direction.
Herbert Simon, who spent a career mapping how organizations actually make decisions, observed that organizations do not make decisions.8 People inside organizations make decisions, shaped by the structures, routines, and shared understanding that surround them. Improve the technology without improving those structures, and you improve the speed and scale at which existing patterns reproduce themselves — for better or worse.
This creates a subtle but consequential bias in management itself. Investment becomes easier to justify than integration. New capabilities become easier to celebrate than better judgment. Over time, organizations become increasingly sophisticated at acquiring technologies while remaining comparatively underdeveloped in understanding how those technologies change the cognitive demands placed on the organization. Technology has never been management's final object. The confusion between the instrument and the object may be the most persistent and expensive mistake in the history of the discipline.
Capability and complexity compound together
Every major technological revolution has produced the same surprise.
As organizations become more capable, they also become more complex. Machines increased production, but they also increased coordination requirements. Information systems increased access to knowledge, but they also increased the volume of decisions requiring interpretation.10 Artificial intelligence dramatically expands individual capability while simultaneously raising the cost — and the difficulty — of organizational coherence.
Technology does not simplify organizational complexity. It amplifies it.
This creates a paradox that sits beneath every wave of enterprise technology disappointment. The more powerful the capabilities an organization deploys, the more it depends on organizational intelligence it may never have developed. The gap between what individuals can now do and what the organization can coherently decide to do with those capabilities grows wider, not narrower, with each technological generation.
It may explain why organizations routinely report individual productivity gains that do not translate into organizational performance gains. The individuals are genuinely more capable. But capability distributed across thousands of specialized people, without the architecture to integrate it, produces something other than intelligence. It produces activity — faster, louder, and more expensive than before.
The object management missed
The two previous articles in this series each performed a specific reframe. The organization became a cognitive system, not merely an economic one. Specialization became the structural reason collective intelligence is necessary, not simply a way to improve efficiency. This article draws the implication that neither of those moves made explicit.
If organizations are cognitive systems, then the true object of management has always been the organization's capacity to think — to perceive, integrate, learn, and decide collectively. Technology has never been that object. It has always been an instrument through which that capacity gets tested, revealed, and either developed or neglected.
Every technological revolution has been, at its core, an invitation to ask a question that management mostly declined to answer. Not: how do we deploy this technology? But: what does this technology reveal about how our organization actually thinks?
A more consequential question
The current AI reckoning — budgets exhausted, returns elusive, governance scrambled — may be the clearest invitation yet to ask that question seriously.
If the limiting factor has never been the technology, and if technology has always been an amplifier of whatever organizational capacity already exists, then the question every leadership team should be asking is not which AI to deploy. It is what kind of organization they are building to use it.
That question leads somewhere most management frameworks have not yet gone — not to the technology, not to the hierarchy, not even to the talent, but to the largely invisible substrate through which people in organizations develop shared understanding, coordinate action, and become capable of thinking together.
Every organization already possesses this substrate. It exists whether leadership recognizes it or not. It shapes every meeting, every decision, every disagreement, and every strategic initiative. Yet because it rarely appears on an organizational chart or a financial statement, it remains largely invisible until moments of extraordinary change expose it.
What that substrate actually consists of — and why the most consequential decisions an organization makes may happen long before anyone enters a meeting room — is where this investigation goes next.
───
Selected References
Enterprise Evidence
1. TechCrunch. The Token Bill Comes Due: Inside the Industry Scramble to Manage AI's Runaway Costs (2026).
2. TechCrunch. Companies Are Scrambling to Stop Employees from Maxing Out AI Budgets with Small Tasks (2026).
3. Gartner. Autonomous Business and AI Layoffs Create Budget Room but Do Not Deliver Returns (2026).
4. Financial Times / Australian Financial Review. Enterprise AI governance and AI operating costs (2026).
5. Business Insider. Enterprise AI adoption, organizational incentives, and AI governance (2026).
6. Oplexa. The AI Inference Cost Crisis (2026).
Foundations of Organizational Thought
7. Hayek, F. A. (1945). The Use of Knowledge in Society.
8. Simon, H. A. (1947). Administrative Behavior.
9. March, J. G., & Simon, H. A. (1958). Organizations.
10. Galbraith, J. R. (1973). Designing Complex Organizations.
11. Weick, K. E. (1995). Sensemaking in Organizations.
12. Coase, R. H. (1937). The Nature of the Firm.
13. Williamson, O. E. (1985). The Economic Institutions of Capitalism.
14. Hackman, R. (2002). Leading Teams.
Explore the OCI series
Previous: Why Organizations Need Intelligence
Next: The Illusion of Alignment
Discuss organizational co-intelligence in your organization.