Everyone Used to Wear a Hat — What History Teaches Us About Agentic AI
For most of Western history, stepping outside without a hat was unthinkable. Not merely unfashionable — socially illegible. The hat was a signal of respectability, profession, and belonging. Then, over roughly a decade in the 1960s, it disappeared almost entirely. Not because hats stopped working. Because the world they expressed stopped existing.
That story is more instructive than it first appears — especially if you are trying to understand what agentic AI is about to do to the modern organisation.
The Hat Was Never Just a Hat
Hats served real purposes: protection from sun, rain, and cold. But by the time they vanished, those purposes had long been supplemented by enclosed automobiles, heated offices, and weatherproof buildings. What kept hat-wearing alive was not function. It was the elaborate social scaffolding built around it — hat checks at restaurants, dress codes, etiquette rules about tipping and doffing, and the silent language of status that different hats communicated.
The hat persisted because it was load-bearing infrastructure for a particular social world. And it collapsed when that social world did — when postwar formality dissolved, when the counterculture rejected the visible grammar of hierarchy, when young men like John F. Kennedy appeared bareheaded and made the old uniform suddenly look ceremonial rather than necessary.
The hat didn’t die because it became useless. It died because the social world it expressed stopped existing.
This is the distinction that matters. The car was the proximate cause — it removed the practical need. But the real transformation was cultural: new assumptions about what formality meant, what status should look like, and what it meant to be a serious person in public life.
Most Organisational Rituals Are Hats
Look honestly at how large organisations actually spend their time. Status meetings that report what everyone already knows. Approval workflows that exist to distribute accountability rather than improve decisions. Slide decks prepared not to inform but to demonstrate that work is happening. Email chains that loop in stakeholders whose primary function is to be looped in.
These are not inefficiencies waiting to be optimised. They are social performances — the organisational equivalent of hat-wearing. They signal seriousness, effort, and belonging. They make the institution recognisable to itself. And like hats, they will not yield to logic alone. They will yield when the conditions that make them meaningful stop existing.
Agentic AI is changing those conditions.
Agents Are the Automobile, Not a Better Horse
The error most organisations are currently making is treating agentic AI as a productivity tool — a faster, cheaper way to do what humans already do. This is the equivalent of strapping an engine to a carriage and calling it a car. The car didn’t make horses slightly less useful. It restructured the entire geography of daily life so comprehensively that the horse’s role became unrecognisable.
When an agent can autonomously gather information, synthesise it across sources, brief stakeholders, draft the follow-up, execute the next step, and flag exceptions for human review — the elaborate human choreography built around those activities loses its structural justification. Not because the people doing it are incompetent, but because the conditions that required that choreography have changed.
The question is not which tasks will agents handle. The question is which conditions of work are being permanently altered — and therefore which rituals, roles, and structures will find themselves, like the hat, suddenly without a world to express.
Six Things the Hat Teaches Us
1. Resistance will come from the infrastructure, not the people
Hat-wearing persisted partly because of the ecosystem around it — hat racks, dress codes, social scripts. Organisations have equivalent load-bearing structures: org charts built around information gatekeeping, job descriptions filled with coordination overhead, performance metrics that measure activity rather than outcomes. These will resist transformation not because they are useful, but because they are how the institution recognises itself as functioning.
2. There will be a JFK moment — and it may be closer than you think
Kennedy didn’t kill the hat alone, but he was a visible signal that the shift was real and sanctioned from the top. Every organisation is currently waiting — consciously or not — for its own bareheaded moment: a leader who visibly works with agents, not just alongside them, and whose results make the old way look ceremonial by comparison. That moment tends to arrive faster than institutions expect, and once it arrives, the old normal looks absurd in retrospect.
3. The technology is the proximate cause. The transformation is cultural
Organisations that treat agentic AI as a tooling question will be perpetually behind. The real questions are civilisational: What does accountability look like when execution is automated? What is human presence in a process actually for? What does a serious professional look like when the choreography is gone? These are not HR questions. They are questions about identity, value, and what it means to work.
4. New expressive surfaces will emerge — and they will be more individual, not less
When hats disappeared, hair became the expressive surface — personal, visible, impossible to standardise. As agents absorb execution and coordination, what humans bring that cannot be agentified becomes the new differentiator: ethical judgment, contextual taste, relational trust, the capacity to ask the question that wasn’t on the agenda. The professionals who will thrive are not those who resist agents. They are those who figure out what they uniquely express when the hat is gone.
5. Governance will lag — and the lag will be costly
Society didn’t immediately build new traffic law, urban planning, and infrastructure for the car. The transition produced decades of costly improvisation. The same is true here. Organisations deploying agents without rethinking accountability structures, oversight models, and decision rights are building motorways with horse-carriage rules. The governance question is not a compliance afterthought. It is the core design challenge.
6. The winners won’t be the fastest movers. They’ll be the clearest thinkers
The first automobiles were unreliable, dangerous, and impractical. The people who shaped the automobile era were not those who bought the first car, but those who understood what kind of world the car made possible — and built for that world. The same will be true of agentic AI. Speed matters less than the quality of the mental model you bring to the transformation.
So — Where Does That Leave You?
The hat was not abandoned overnight, and it was not abandoned by everyone at once. There were people who kept wearing hats long into the hatless era — some out of genuine preference, some out of habit, some because their particular world still required it. There will always be contexts where the old forms persist and serve real purposes.
But for most organisations, in most sectors, the question is no longer whether agentic AI will fundamentally alter the conditions of work. It already is. The question is whether you are building for the world that is emerging — or defending the world that is passing.
History is not kind to hat defenders. Not because they were foolish, but because they were loyal to the wrong thing: to the artefact, rather than to the purpose the artefact once served.
The hat is the workflow. The agents are the car. The only real question is what you are going to do with your hair — what you will bring, distinctly and irreplaceably, when the choreography is finally someone else’s job.
#agenticAI #AITransformation #digitalLeadership #enterpriseAI #futureOfWork #organisationalChange