The Ratchet

In the AI Story

Stage one is sanctioned experimentation. Management provides AI tools to a subset of the workforce with guidelines, use-case restrictions, and review processes. This stage resembles what March would recognize as genuine exploration: an investment in uncertain technology with explicit understanding that the experiment may fail. Stage two is where the ratchet engages. Individual practitioners, working within experimental bounds, discover the tool is more capable than anticipated. Discoveries are shared informally — in Slack, over lunch, in the hallway conversation that has always been the most efficient information channel. Practitioners do not think of them as discoveries but as small efficiencies, life hacks, tricks of the trade, worth sharing but not worth reporting.

The aggregation of small efficiencies produces something worth reporting to management, and by the time it is reported, it is no longer a discovery but a fait accompli. Documentation is now AI-generated. Reports are AI-drafted. Test cases are AI-produced. No one decided that these activities would be transferred to the AI. Each practitioner made a local decision — this task, this time, this tool — and the accumulation produced a systemic transfer of organizational capability from humans to machines.

Stage three is normalization. The AI-assisted way of working becomes default. New employees are trained on the AI-augmented workflow. Process documents update. The non-AI way persists in institutional memory but fades from institutional practice. A developer who wanted to write code without AI could in principle do so; in practice, the development environment, timeline expectations, and performance benchmarks have all been calibrated to the AI-augmented workflow. Working without AI is not prohibited, merely impossible — in the way that driving a horse to work in a city designed for automobiles is not prohibited but is merely impossible. Stage four is dependency: the organization cannot function without the AI tools, not because they have been designated as critical infrastructure but because the human capabilities they replaced have atrophied through disuse.

The Trivandrum training compressed the entire sequence into five days. By Friday, the transformation was described as 'measurable, repeatable reality.' What is absent from the account is a decision. At no point did anyone make a deliberate choice to abandon the old way of working. The old way was not abandoned; it was rendered obsolete by the accumulation of individual discoveries that made continuation irrational. The ratchet engaged through the mechanism March described: not through decision but through the accumulated weight of adaptations that no one tracked.

Origin

The ratchet concept synthesizes March's work on routines (with Cyert in A Behavioral Theory of the Firm), his observations about how organizational change actually occurs versus how it is narrated, and his insistence that the official decision process and the actual decision process operate in parallel, connected only loosely. The AI-era extension draws on empirical observation of how organizations have actually adopted AI tools — a process systematically different from the rational-planning model that AI strategy documents describe.

The ratchet's irreversibility derives from an asymmetry between the costs of adoption and the costs of reversal that March identified in his work on path dependence. Adoption is cheap because it piggybacks on existing motivation — the practitioner wants to work faster, the manager wants to hit the quarterly target. Reversal is expensive because it opposes existing motivation — it asks practitioners to be slower, managers to accept lower output, organizations to invest in capabilities they do not currently need.

Key Ideas

Four stages. Sanctioned experimentation gives way to informal adoption, normalization, and dependency, with no stage-gate review at any transition.

Local rationality, systemic unexamination. Each individual decision is rational; the aggregate transformation is never evaluated at the system level.

One-directional motion. The ratchet engages through motivated adoption and resists reversal that would require counter-motivation no learning signal supports.

Discourse outruns experience. The ratchet moves faster than the organization's capacity to evaluate what the ratchet is doing, producing retrospective rationalization rather than prospective evaluation.

No author, no record. The ratchet leaves no trace in organizational records designed to capture decisions, because it never produced one.

Debates & Critiques

Critics argue that the ratchet framing understates the capacity of thoughtful leadership to slow or redirect the transition. Defenders respond that the thoughtful leaders are themselves embedded in the same learning systems that drive the ratchet, and that their attempts to redirect tend to fail not through lack of commitment but because the structural forces producing the ratchet are stronger than individual will. The framework does not deny that leadership matters; it specifies what leadership must do to matter — construct structural protections against the drift, rather than attempting to resist the drift through individual vigilance alone.