How AI is really changing business, strategy, and organisations

We tracked code quality metrics across 14 engineering teams using AI coding agents and found that while velocity increased by 40%, architectural coherence declined by 28% and cross-module defects tripled. Teams are measuring the wrong thing.

Every AI agent session starts from zero. The architectural decisions from six months ago, the module boundaries drawn for specific reasons, the bug fixed three times in three different ways. We examined how knowledge loss compounds across agent sessions and what it costs.

The dominant workflow for AI coding agents is prompt-and-pray: describe what you want, hope the output is correct. We studied 11 engineering teams and found that teams using structured design canvases before agent execution produced 62% fewer cross-module defects and 3x faster review cycles.

Code review was designed for human-speed output. When agents produce ten times the volume, manual review becomes theatre. We found that reviewers catch 23% of defects in agent-generated code versus 61% in human-written code, and the gap widens as volume increases.

For decades, organisational intent survived in people. The architect who remembered why. The strategist who held the thread. AI is removing the carrier, and what's left are static artefacts that were already inadequate. We argue that intent must become a living system state.

We studied how five organisations retain and compound institutional knowledge through AI-assisted workflows. The gap between those with persistent knowledge systems and those without widened with every quarter. By month nine, structured teams were operating at fundamentally different capability levels.

We worked with three organisations to replace their annual strategy cycle with a continuous system of falsifiable bets wired to live evidence. Within six months, two had identified and killed strategic initiatives that quarterly reviews had rated as 'on track' for over a year.

We traced 16 strategic signals from operational reality to executive reporting across two organisations. On average, 67% of signal content was lost, transformed, or delayed by more than 90 days. The quarterly review is structurally incapable of seeing what matters.

We measured the gap between executive confidence in strategic execution and actual evidence of strategic progress across four enterprises. The average confidence-evidence gap was 41 percentage points. Every executive team was substantially more confident than their data warranted.

A detailed case analysis of how initiatives labelled 'strategic' become untethered from the outcomes they claim to serve, and why every standard reporting mechanism missed it.

How to follow a single signal (a customer complaint, a risk flag, a market insight) through an organisation and measure exactly where and how it degrades into uselessness.

The prevailing assumption is that autonomous AI agents are the future of enterprise automation. We argue that auditable, deterministic pipelines with human validation produce better outcomes, and explain why the distinction matters.

The dominant mental model treats strategy as a plan to be executed faithfully. We propose an alternative that changes what you measure, what you track, and what you do when things aren't working.

OKRs were designed to create alignment between strategy and execution. In practice, they often do the opposite, creating an illusion of alignment that masks deeper disconnection. We examine how this happens and what to do about it.

Every AI deployment changes the organisational structure, not on the org chart, but in the actual flow of information, decisions, and power. We mapped these invisible restructurings across eight enterprises and found that none had been anticipated, planned for, or even noticed.

Every AI business case models the direct effects: cost reduction, speed improvement, accuracy gains. Almost none model the second-order effects: how AI changes team dynamics, shifts power structures, alters information flow, and reshapes organisational boundaries.

The consensus is that generative AI will automate routine tasks. We argue the bigger disruption is structural: AI is compressing the middle management layer that serves as the primary translation mechanism between strategy and execution.

Technical debt has a well-understood cost structure. Strategic debt, the accumulated weight of deferred organisational decisions, is more expensive and far less visible. We found $23M in annual cost attributable to decisions that were never made.