The Rise of the Synthetic Organisation Part 2

There is a conversation that does not happen in most boardrooms.

Not because leaders lack intelligence, or courage, or access to data.

But because it requires someone to say the thing nobody wants to put in the board paper: every system we have built — governance, incentives, culture, structure — was designed to protect and improve what exists. Not to replace it.

The organisational immune system.

Decades of accumulated compliance frameworks, approval chains, performance management systems, and job architectures that have learned to identify transformation as a foreign body — and neutralise it. Not through malice. Through design.

This article is about what it takes to overcome it.

The synthetic organisation has answered the question most organisations are still avoiding: what work is AI for, and what work belongs to humans? The low-judgment, repetitive, administrative work — gone. The people inside it do what only people can do. And the competitive gap between this organisation and the one still running on the old design widens every quarter.

A word on what this means for jobs. The elimination of low-judgment work is not consequence-free, and the people affected by that transition deserve more than a line about reskilling. This series will address the human impact directly — it is too important to treat as a footnote to a strategy framework.

The Immune System Is Still Winning

Two and a half decades in workforce and technology design give you a particular vantage point on AI. Not every technology transition rewrites the game. Most don't. This one does. And the question it forces — what is the best way for humans and AI to work together, and why is almost nobody actually designing for it — has dominated the last five years of my work. And it is not going away.

Most of the advice circulating right now frames this as a technology transformation. It isn't. It is a workforce transformation. The organisations that get that distinction right are the ones that will build the synthetic organisation. The ones that don't will spend a fortune on AI and wonder why nothing changed.

Workforce is not a function. It is the entire operating model — every role, every process, every decision about who does what and why. When AI rewrites the answer to that question, it rewrites the organisation. That is not an HR agenda. It is a CEO agenda — requiring time away from business-as-usual, structured, protected, and owned at the top. The immune system will consume every transformation initiative that competes for attention inside the normal rhythm of the business.

BCG's 2026 AI Radar survey of 640 CEOs found that half believe their job stability depends on getting AI right this year. The same survey found that 60% have deliberately slowed implementation. That is not a contradiction. That is rational paralysis — the product of an incentive system that makes individual caution logical and collective progress structurally improbable.

The immune system, left unreformed, produces exactly these numbers. Every time.

The Matrix Moment: Kodak, Fujifilm, and the Choice Every Leader Faces

In 1975, a Kodak engineer named Steve Sasson built the world's first digital camera inside the company's R&D lab. Management's response was essentially: interesting — don't tell anyone.

This was not stupidity. Film margins were close to 70%. The rational calculation for every executive whose compensation was tied to film division performance was obvious: slow this down. Protect what's working. Individually logical. Collectively — organisational armageddon.

Fujifilm faced the same disruption at the same time. Demand for photographic film collapsed to one-tenth of its peak between 2000 and 2010. Both companies saw it coming. Both had profitable businesses to protect. Both had leadership teams shaped by decades of analogue success.

Fujifilm chose to see the world as it was, not as it had been. That sounds simple. It is not. Seeing clearly — really clearly, ahead of the data, ahead of the consensus, ahead of the comfort — requires a particular kind of bravery. The willingness to name what is true before it is safe to say so. Fujifilm mapped its entire technology portfolio — identifying more than 70 capabilities transferable from its film chemistry heritage (Komori, Innovating Out of Crisis, 2015) — and systematically rebuilt it around healthcare, cosmetics, and semiconductor materials. Today, it generates record revenue.

Kodak's management chose the blue pill. Fujifilm chose the red one. The technology was identical. The result was not.

Fujifilm's CEO Shigetaka Komori later described five ingredients he believes separate leaders who pull off transformation from those who don't.

Heart first — without genuine care for people, nobody follows you into the unknown.

Then intuition — the ability to see what is coming before the data confirms it.

Then courage — the willingness to act on that reading while everyone around you is still defending the status quo.

Balance and decisiveness follow.

They are what allowed Komori to invest $400 million in a new research facility while simultaneously cutting 5,000 jobs — and have people follow him. He was not reacting to the data. He was acting ahead of it.

Inside most large organisations right now, the choice is already here. The red pill or the blue pill. See the world as it is and act — or protect what exists and hope the disruption is slower than it looks.

The People Building This Are Not Equivocal

Sam Altman (blog post, January 2025): AI agents are now doing real cognitive work. By the end of 2026, AI systems will begin producing genuinely novel scientific insights.

Dario Amodei (Machines of Loving Grace, 2024): AI will disrupt 50% of entry-level white-collar roles within one to five years. Not the next decade. This one.

These are not hedged forecasts from commentators. They are the stated convictions of the people closest to what is being built. Treating them as distant hypotheticals is the contemporary equivalent of Kodak's management telling Steve Sasson to keep quiet about the camera.

Five Reasons The Immune System Keeps Winning

The first is the board knowledge gap. Two-thirds of board directors report limited to no AI knowledge. Nearly one in three say it does not appear on their agenda (KPMG Board Survey, 2024). MIT Sloan research found organisations with AI-savvy boards outperform peers by 10.9 percentage points in return on equity. This isn't a governance hygiene issue. That is a valuation gap — and a CEO who wants to move fast has no cover without it.

The second is the incentive trap. The median S&P 500 CEO tenure is under five years. A genuine AI transformation takes three to five years to yield its most significant returns. The arithmetic is brutal: the leader who initiates the transformation is unlikely to be the one who reaps it, but will absorb every quarter of disruption along the way, while compensation is tied to the short-term metrics, the transformation temporarily worsens. Every manager in the organisation is running the same calculation.

The third is the pilot trap. Eighty-eight percent of organisations have AI running somewhere (McKinsey, State of AI, 2024). Most have become managed experiments — contained enough to be defensible, small enough to be deniable. And many have made a more specific mistake: they have deployed a productivity tool — Microsoft Copilot, Google Gemini, a ChatGPT enterprise licence — and called it an AI transformation. Copilots assist. They wait for a prompt, generate a response, and return control to the human. They are valuable. They are not transformative at scale. The pilot trap is comfortable. It produces progress reports without producing transformation.

The fourth is the middle management layer — the immune system's engine room. If AI can automate 40% of a manager's team, championing that transformation means volunteering to shrink their own budget and diminish the metrics by which they are evaluated. They are not being obstructionist. They are being logical. Multiply that calculation across every middle manager in a large organisation, and you do not get resistance. You get a system producing exactly the outcomes its incentives predict.

The fifth is the communication vacuum. Shareholder AI proposals have grown fourfold in recent years (Harvard Law School Forum on Corporate Governance, 2024). Employees, investors, and regulators are forming views about an organisation's AI posture — often in the complete absence of a signal from leadership. The vacuum fills with anxiety, speculation, and the assumption that silence means no strategy, or a strategy that cannot withstand scrutiny.

Six Moves To Overcome The Immune System

The immune system is not irrational. Job architecture, skills taxonomies, governance frameworks, risk management — these are the mechanisms organisations use to hold strategy in place. They work brilliantly for stability. The problem is that stability is precisely the wrong design for a moment when the fundamental question of who, or what, does the work is being rewritten.

Overcoming the immune system does not mean burning down the governance structures. It means asking a harder question: are these defences protecting the organisation, or protecting the people inside it from having to change?

It requires deliberate structural intervention. Not another pilot, not another town hall, not another values poster. A different kind of leadership architecture entirely.

Name your AI posture. Most organisations have an implicit AI posture they have never made explicit. The vacuum fills with competing interpretations, conflicting pilots, and a leadership team that cannot align because they are not working from the same map. Fujifilm named its bets early, identified the specific domains where its existing capabilities gave it an asymmetric advantage, and committed. An organisation that cannot name its AI posture cannot govern it, fund it properly, or hold anyone accountable for it.

Build the coalition. Two-thirds of board directors report limited to no AI knowledge. MIT research found organisations with AI-savvy boards outperform peers by 10.9 percentage points in return on equity — a valuation gap, not a governance hygiene issue. The intervention that works is directors using the tools themselves, engaging directly with AI-native competitors, and sitting with competitive intelligence rather than being briefed on it.

Redesigning the incentive architecture. The immune system's power source is incentives. Until you change what gets measured and rewarded, the rational individual calculation for every manager remains: protect what I have. McKinsey is direct: people allocation, career structures, performance management, and reporting systems all need to be redesigned. The mechanism is a parallel AI transformation scorecard — visible at the board level, tied to executive compensation, measuring pilot-to-production conversion and strategic value created. BCG research found employee confidence in AI rises from 15% to 55% when they experience genuine leadership commitment (BCG, Winning with AI, 2024). When Jensen Huang discovered some Nvidia managers were discouraging AI use, his response at an all-hands was: "Are you insane?" The incentive architecture must make that response the norm.

Deploy the Black Belts. The middle managers who are currently the immune system's most effective agents of resistance are also the organisation's highest-leverage change agents, if their mandate and incentives are reoriented. The Six Sigma Black Belt model is the template — a senior practitioner embedded in each business unit, accountable for local transformation outcomes, reporting to a central Centre of Excellence for standards and to the business unit leader for results.

What the Black Belt actually builds is not a Copilot rollout. What creates structural, compounding value is an AI mesh — a connected network of agents, each scoped to a specific use case and workflow, orchestrated across functions, and governed for performance over time. MIT research shows that organisations that move from isolated AI tools to orchestrated agent meshes generate returns that are an order of magnitude greater than those of organisations running point solutions.

This is exactly what a Black Belt does. Identify the specific use cases in their domain where an AI agent can execute autonomously. Define the workflow. Establish data access, guardrails, and human-in-the-loop thresholds. Measure business outcomes — not usage metrics. Monitor for drift and failure modes. Improve iteratively. Then connect it to the adjacent agent in the next workflow, building the mesh outward from a foundation of validated, governed deployments. This is deliberate, disciplined work.

Govern AI. Fewer than one in four organisations have board-approved AI governance policies (KPMG AI Governance Survey, 2024) — a governance vacuum at the most consequential point of strategic risk, including the risk of not moving fast enough. A board-approved policy covering risk appetite and escalation protocols; an executive owner with direct board access; a risk framework that models the competitive cost of inaction; quarterly AI metrics alongside financial ones. This is a signal that the commitment is real.

Communicate like a founder. The CEO's dilemma cannot be resolved by governance frameworks alone. It resolves when the CEO makes a public, irreversible commitment — not to AI in the abstract, but to the specific transformation of this organisation, on this timeline, with their own credibility on the line. The commitment reaches five audiences simultaneously: the board hears the risk of inaction; investors hear the return on commitment; lenders hear the covenant confidence; employees hear the organisation being built; customers hear the value they will receive. The leadership act is making all five visible at once — and making the direction of travel impossible to walk back.

The Platform Is Still Burning

The immune system is not a metaphor for inevitability. It is a description of a structural problem with structural solutions. The organisations that close the gap will not do so through ambition. They will do so through the deliberate work of the six moves above — named posture, board coalition, redesigned incentives, Black Belts building the mesh, governed AI, and a founder-level commitment that makes the direction of travel irreversible.

Morpheus didn't promise Neo it would be easy. He promised it would be real.

There is a question I ask in boardrooms that tends to land in silence. When voice calls became free — when Skype and WhatsApp made the telcos' core product worth nothing - the companies that survived anticipated the change and stopped asking how to protect voice revenue. They asked a different question: what business are we actually in? The ones that answered honestly rebuilt around connectivity and data infrastructure. The ones that didn't are now commoditised pipes — some still standing, but strategically irrelevant. When cognitive work becomes automated — when the tasks your current workforce performs can be done faster, cheaper, and better by a redesigned human-AI system — what business are you in?