Introduction

AI’s most consequential capability is not generation but reconciliation: the ability to read strategy, architecture, and code together and surface the gap between narrative and reality. For most organisations, the synthesis collapses into contradiction.

By now, it is clear that AI can generate content; most executives use it daily. This book is not about that capability but about AI's ability to read and reconcile, and the tremendous effect that capability is set to have on competition, operating models, and the people working inside them.

When a machine can read what you say alongside what your systems do, the gap between narrative and reality becomes measurable. A strategy document can be compared against the code that must implement it. An architecture reference model can be compared against what is actually running. A budget line can be compared against the operational outcomes it was meant to produce. A portfolio label can be compared against the services that supposedly comprise it. An investment case can be compared against the incident costs it was meant to prevent.

Once strategy documents, architecture models, budget lines, portfolio labels, and code can be read together, the organisation's claims become testable. Structural honesty becomes measurable rather than aspirational.

Reading and reconciliation are only the beginning. Organisations are deploying autonomous agents to act within their systems, and those agents require the same structural preconditions: processes defined explicitly enough to follow, boundaries clear enough to respect, ownership unambiguous enough to navigate. An agent deployed into a structurally incoherent organisation either fails at every boundary where a human previously compensated through social negotiation, or reproduces the incoherence at machine speed.

The book uses a specific term: software-dependent corporate. It refers to large, established organisations whose products and operations depend critically on software, but whose governance structures were designed for a world in which software was a support function.

Banks, insurers, retailers, logistics firms, airlines, public sector bodies, and older companies that describe themselves as technology companies but still operate with inherited governance assumptions all fall into this category. Between roughly 2008 and 2012, software stopped being infrastructure and became the primary medium through which these companies create and deliver value. The technology changed, but the governance structures, budget cycles, and power structures did not, and the misalignment has been compounding ever since.

...

See also: Full contents · Preview chapters · Illusions of Work