Executive take
Quick answer
This response is based on the framing and McKinsey's known public positions; the full article was not reviewed in detail. McKinsey's piece is positioned as a manifesto - a declared position meant to move leaders from hesitation to action. The likely argument: leaders who treat AI as an IT project will lose to those who treat it as an operating model question. That is a reasonable position. It is also the position McKinsey has been refining for several years. The manifesto tends to favor large-scale, top-down transformation programs. That doesn't make the argument wrong, but leaders should read it with an eye toward what is actionable for their specific situation.
Perspective
Business leader
Why this matters for this role
What this role should do
Watchouts
The thesis
This response is based on the framing and McKinsey's known public positions; the full article was not reviewed in detail. McKinsey's piece is positioned as a manifesto - a declared position meant to move leaders from hesitation to action. The likely argument: leaders who treat AI as an IT project will lose to those who treat it as an operating model question. That is a reasonable position. It is also the position McKinsey has been refining for several years. The manifesto tends to favor large-scale, top-down transformation programs. That doesn't make the argument wrong, but leaders should read it with an eye toward what is actionable for their specific situation.
What the manifesto framing gets right
The instinct behind a manifesto is sound: most organisations are still treating AI as a collection of point tools rather than a shift in how work gets structured. Deploying a chatbot on top of an unchanged process isn't transformation - it's decoration. The pressure to think at the operating model level is real. Leaders who aren't asking structural questions about where AI changes accountability, judgment, and workflow are falling behind those who are. McKinsey is also right that culture and leadership posture matter as much as the technology choices. Teams that are waiting for a perfect understanding of AI before starting will find themselves at a disadvantage.
Where the manifesto framing misleads
The problem with manifestos is that they are written for the organisation that has already decided to move - and they tend to propose comprehensive programs when most leadership teams actually need a smaller, faster win. Most organisations do not need an AI transformation. They need two or three AI deployments that work well enough to build organisational confidence, surface real friction points, and demonstrate measurable return. The 80/20 principle applies here hard: a handful of high-frequency, well-scoped workflows - document summarisation, first-draft generation, structured data extraction, meeting note synthesis - can deliver much of the early value without a full transformation program. Start with one workflow. Measure time saved or quality improved. Use that evidence to inform the next step.
The 80/20 version of AI strategy
Rather than asking 'how do we transform?', the more productive leadership question is: 'Which three workflows, if AI-assisted, would reduce the most friction for the most people in the next 90 days?' That question is answerable. It produces a short list. It leads to a pilot, not a program. Start with tasks that are high-volume, well-defined, and currently slow - not because they are glamorous, but because they are the ones where success is visible and failure is recoverable. Internal report drafting, meeting summarisation, first-pass contract review, and data extraction from unstructured documents have shown early positive results in limited deployments. Independent validation remains thin, so treat these as directional signals rather than guaranteed outcomes.
What it means for leaders
Take the structural framing seriously: AI is an operating model question, not an IT procurement question. Leaders who own that distinction will make better decisions about where to invest and where to wait. Set aside the transformation urgency if it is being used to justify a large, complex, multi-year program before the organisation has demonstrated it can deploy even one AI workflow reliably. The pressure to 'move at scale' before 'moving at all' is where the most expensive mistakes are being made. The McKinsey manifesto is a useful provocation. It is not a plan. The plan has to be built from the ground up, based on the organisation's actual workflows and constraints.
Risks to watch
The biggest risk is treating any manifesto as a blueprint. Organisational context matters more than methodological frameworks. A second risk: over-rotating to one vendor's AI stack because it aligns with the manifesto's consulting model. Vendor lock-in and shelfware remain common pitfalls. Finally, there is the risk of analysis paralysis - spending so long on strategy that competitors who simply start experimenting gain an edge. Mitigate by running short, time-bound pilots that test specific workflows and generate cost or time data within a quarter.
Reader signal
Was this useful?
Reader feedback
Help tune future briefings
Related reading