The Fix

Chapter 15: Where This Fails

Six adversarial cases stated as forcefully as possible and examined for what survives. The honest limits of the model, including the most dangerous failure mode: governance capture that produces clean dashboards while dysfunction continues.

The preceding two chapters established the structural preconditions: the P&L architecture must align ownership of outcomes with ownership of execution, and the unit itself has been described: its composition, boundaries, contracts, and interactions. Before proceeding to implementation, the strongest objections to the thesis deserve examination on their own terms.

What follows is not a summary of concerns raised by sceptical readers but an attempt to break the argument from the inside, using the book's own evidence and mechanisms. Each case is stated as forcefully as possible, then examined for what survives. Where the objection wins, the text says so. This is not a common feature of business books, most of which prefer to leave their most vulnerable arguments unexamined on the reasonable assumption that the reader will be too polite to press the point.

The attack. The board pack vignette earlier in the book is the weapon. An AI assistant produced a technology summary that was individually defensible in every fact and collectively fictitious. It did so in thirty minutes. It sounded more authoritative than anything a human would have written. The CTO marked up three corrections and sent it back, knowing the corrections would not land before the board pack shipped.

Now scale this. Every team in the organisation has access to the same generator. Product requirements documents reference architectures that have been decommissioned. Quarterly reviews connect strategy to outcomes through plausible-sounding chains that no one verifies. Process definitions are generated from templates, internally consistent, formatted correctly, and bearing no relationship to how the system actually behaves. The generator layer scales illusion faster than any human verification process can contain it.

Push further. Organisations discover that AI-generated process definitions pass the AI synthesis checks described in this book. The artefacts are syntactically correct, structurally complete, and entirely synthetic. A claims process definition describes states, transitions, failure modes, and contract dependencies that match the expected schema. The synthesis compares the definition against the codebase and finds no contradictions, because the definition was reverse-engineered from the code by the same AI that will later audit it. The artefact looks like clarity. It functions as camouflage.

This is the strongest version of the attack: the generator produces artefacts that satisfy the reader, and the organisation achieves the appearance of structural coherence without the substance.

What survives. The attack succeeds against document-based verification and fails against runtime reconciliation.

The generator/reader asymmetry described earlier in the book is architectural, not documentary. When this book describes binding artefacts, it means something specific: the process definition constrains deployment, not merely describes it. A binding process definition is version-controlled, referenced in the governance charter, and deviation from it triggers a synthesis flag. The mechanism is reconciliation against system telemetry, not document review.

...

See also: Full contents · Preview chapters · Illusions in the Boardroom