Consultants at McKinsey do not publish a second edition of a book they released barely two years ago every day. The fact that they did says something. In April 2026, "Rewired" appeared in a revised version, with interviews and additions based on what large organisations actually learnt over the past years. The conclusions are surprisingly unsurprising. And that is precisely the point.
Because what Kate Smaje and her colleagues put to paper is not the story many technology vendors would like to hear. It is a story about people, patience and choices, not about platforms, licences and scale.
What it says
The core of the book, now confirmed by two years of practical experience at large organisations, is this: AI transformation delivers real returns, but only for those who approach it correctly. The figures are concrete. An EBITDA uplift of 20 per cent is achievable. The payback period is one to two years. For every euro invested, three euros can come back. But, and this is the essence, only at organisations that focus on the economic leverage points. Not at those who start everywhere at once.
McKinsey calls this the difference between "scaling AI" and "spreading AI". Rolling out is not the same as distributing across as many use cases as possible. The organisations that win choose two or three processes where AI genuinely makes a difference to the outcome, and go all in. Organisations that start everywhere at once, without a clear sense of priority, lose focus and returns. That is a different matter from training employees broadly in AI tools and AI literacy, which runs alongside the strategic bets and lays the foundation for them.
"Every AI transformation is, at its heart, a people transformation." That sentence from Kate Smaje should hang in every boardroom. Not as an inspirational poster, but as a warning. Because most organisations still treat AI as a technology project. They buy a licence, roll out a tool, and wait for the productivity gain. That productivity gain does not come. Or not enough. Not because the technology falls short, but because the people are not ready for it, do not have ownership of it and do not understand what is expected of them.
What this means for organisations
McKinsey writes from the perspective of large multinationals. The principles hold at any scale, but not every intervention delivers the same return. The distinction between what works broadly and what delivers deep returns is precisely where most organisations go wrong.
The second edition of "Rewired" introduces a concept that was not in the first edition: LQ, the Learning Quotient. Alongside IQ and EQ, there is now a third dimension that determines how agile an organisation is in a changing environment. The ability to learn quickly, adapt and start again. McKinsey argues that the fastest-learning organisations are the winners, not the largest, not the most technologically equipped.
This is not a metaphor. It is an observation based on what concretely happened over the past two years at organisations that did and did not integrate AI well.
Broad and deep, not broad or deep. McKinsey describes two levels that reinforce each other, but that are too often treated as alternatives. The first level is broad AI literacy: every employee raising their personal productivity through AI, every department with a Champion who pulls that along. The second level is targeted strategic investment: multidisciplinary teams tackling the two or three processes where AI carries the most weight. Both levels are indispensable. Those who only go broad miss the returns that leverage points deliver. Those who only go deep on strategic bets without a foundation in the wider organisation lose the embedding and the buy-in. Robert Levin puts it this way: "A rewired organization is one that has graduated to truly impactful, distributed innovation across the organization. Each manager understands their role, has the right mindset, and is able to innovate with technology and improve their portion of the business." That is not an elite transformation for a handful of teams. That is an organisation that learns and innovates at every level. Kate Smaje adds, on LQ, that it is not just learning for yourself, but "learning for the collective — being able to inspire others and bring the organization along." Breadth makes depth possible, and the strategic bets point to where Champions are needed most.
Focus on leverage points, not on coverage. McKinsey explicitly advises organisations against rolling out AI broadly as a first step. Choose a process where AI makes a demonstrable difference to the end result, measure it, and scale from there. "Don't let perfect get in the way of good enough", appears literally in the second edition. Learning quickly is more important than planning perfectly. The organisations that learn the most in the first year have the greatest lead after three.
The N-2 leader is the key figure. McKinsey identifies a specific role as crucial: the manager at the second or third level below the top, who owns an end-to-end process and embeds technology within it. Not the CTO who maps out the strategy. Not the user learning a tool. The person in between, who knows the domain inside out and understands how technology can strengthen that domain. Whoever has that person in their organisation and gives them the space and resources has already won half the transformation.
Transformation is not a project. "Rewired is a muscle you're constantly honing", writes Smaje. AI adoption has no end date. There is no moment at which an organisation is "done". Organisations that understand this build continuous learning capacity rather than one-off training programmes. That requires a different mindset from management: not "when will this be finished", but "how do we keep growing in this".
Data is specific, not generic. One of the practical insights from the second edition is that generic AI tools deliver far less than tools working on an organisation's specific data. McKinsey advises investing in domain-specific datasets, not in the broadest possible AI application. Those who understand their own data and connect that data to the right tools consistently achieve more return than those who buy what the vendor recommends.
The real lesson
What McKinsey puts to paper in "Rewired" is not news for everyone. Those who have been involved in AI adoption programmes at mid-sized organisations over the past two years will recognise each of these insights. The large technology vendors sell coverage. They want you to roll out as many tools as possible, because that is how their model works. But the organisations that genuinely achieve returns do the opposite.
They do not start everywhere at once. They choose where it counts. They invest in the people who carry every level of the organisation, from the employee who approaches daily tasks differently to the multidisciplinary team redesigning a strategic core process. Broad enough to land, deep enough to deliver returns.
The McKinsey report now gives that approach a solid external foundation. The numbers are there. The principles are validated. What remains is the question of which organisations actually do something with them.
The technology is available. The knowledge too. What makes the difference is ownership. Not at a consultant or a vendor, but with the people who do the work every day, at every level of the organisation.