The Post-Managerial Organisation: What Remains When Intelligence Gets Cheap

In 1937, Ronald Coase asked why firms exist at all. His answer: markets are expensive to use. Searching, negotiating, and enforcing every transaction costs money, so we build organisations to internalise that friction. Herbert Simon later added the human constraint — managers exist because no single mind can process everything, so we stack layers of people, each compressing information for the layer above. The pyramid, in other words, was never sacred. It was a workaround for two scarcities: coordination and cognition.

Both scarcities are collapsing. When AI systems can summarise, monitor, forecast, allocate, and reconcile at near-zero marginal cost, the informational rationale for middle management thins out. This does not mean hierarchies vanish — accountability still needs an address — but it means the pyramid stops being the default answer. The question for the next twenty years is not "flat or tall?" It is: what do organisations do when execution is abundant and something else becomes the binding constraint?

India has already prototyped one answer. NPCI, the institution behind UPI, employs a tiny fraction of the workforce of the banking system it transformed, yet it coordinates billions of monthly transactions across hundreds of competing banks and fintechs. It did this not by managing participants but by publishing a protocol — shared rails, clear rules, open participation. ONDC is attempting the same for commerce. These are organisations whose primary output is not a product but a grammar that lets thousands of independent actors cooperate without a supervisor. Call it the protocol organisation: a small, high-trust core stewarding rules, surrounded by a large, permissionless edge doing the work. Mumbai's dabbawalas ran a low-tech version of this for a century — near-zero management, near-perfect delivery — proof that coordination through shared code predates software.

If coordination is no longer scarce, what is? Three things, and they are related.

Judgment: AI generates options; someone must own the choice, especially the irreversible ones. The manager of 2046 will look less like a supervisor and more like an editor — setting context, defining what "good" means, and taking responsibility for calls machines should not make alone. Management shrinks in headcount but deepens in consequence.

Trust: When anyone can spin up capability, the durable differentiator is whether stakeholders believe you will behave well when it is costly to do so. The Tata group's endurance through 150 years of upheaval owes less to any structure — which has been reorganised repeatedly — than to a reputation that lets it attract talent, capital, and government partnership on favourable terms. Trust is the one asset that compounds across decades and cannot be rented. Future balance sheets will not list it, but future valuations will price it.

Purpose: In a protocol organisation, purpose does the work rules used to do. You cannot supervise a permissionless edge; you can only make its participants want the same outcome. This is why "purpose-driven" is shifting from an HR slogan to an architectural requirement. An organisation without a legible purpose cannot decentralise safely — it can only decentralise chaotically.

The institutions that endure, then, will look strange by today's standards: small cores of perhaps a few hundred people holding judgment, trust, and purpose; vast networked edges of partners, gig specialists, and AI agents holding execution; and hierarchies that survive only where accountability genuinely requires them — safety, capital, law. The Indian opportunity is real. A country that leapfrogged branch banking with UPI and landlines with mobile can leapfrog the industrial-era corporation itself, building protocol institutions for logistics, healthcare, and skilling before legacy structures calcify.

Coase's firms were answers to the cost of markets. The next generation of institutions will be answers to a different question: when machines can do almost everything, what must humans still be trusted to hold? Organisations that answer that clearly will not merely adapt to the future. They will be its infrastructure.