The Only Competitive Advantage Left that You Can’t Buy

By Amrit Sandhar, Founder & CEO, &Evolve

A capability you can buy is not a competitive advantage. It is an entry fee. Goldman Sachs estimates that around a trillion dollars will go into generative AI infrastructure over the coming years, and in 2025 Meta, Amazon, Alphabet and Microsoft planned more than $320 billion of it between them. Beneath them, thousands of organisations are buying the same models, on the same clouds, at the same price. The spending is real. The differentiation is not. So what happens when every competitor acquires the same capability in the same year? Not an edge. A new and expensive baseline.

Forty years ago, Jay Barney asked whether organisational culture could be a source of sustained competitive advantage. He later gave us the test still taught in every strategy classroom: a resource confers lasting advantage only if it is valuable, rare, hard to imitate and hard to substitute. AI is extraordinarily valuable, and fails the other three by design – its entire commercial logic is that anyone can have it by Friday. Culture passes, for the unglamorous reason that nobody can buy yours, poach it, or reverse-engineer it from the outside.

None of which means the AI is wasted. It means something more uncomfortable: returns will vary enormously between organisations running identical technology, and the variance will not be technical. A much-discussed MIT report last year found the overwhelming majority of enterprise AI pilots produced no measurable effect on the P&L. Its method has been argued over, fairly, but the diagnosis survives: these were failures of learning and integration, not model quality. The tool worked. The organisation around it did not.

Which makes it an operating problem, not a procurement one and that puts it on your desk as the CEO, not your technology director’s.

AI runs on an organisation’s information. Not its data lake, its information: what people know, notice, doubt and are willing to say out loud. We have known for four decades where organisations leak. In 1985, Stasser and Titus demonstrated the hidden profile: groups spend their discussion on what everyone already knows and skate past what only one person holds – even when that unshared piece is the very thing that would change the decision. A meta-analysis of 72 studies and nearly 5,000 teams confirmed it: how well information moves predicts how well teams perform.

So the binding constraint on most organisations is not compute. It is whether the person who knows the thing says the thing. Which makes this a question about managers still the most critical and most neglected lever we have.

Picture an ordinary Tuesday. Twelve people, an hour, a decision to make. The most senior person opens by setting out where they have landed, reasonably, to save time. Two others agree. A team leader four rows down the structure has watched this go wrong before, in a depot nobody in the room has visited, and does the arithmetic we all do: how long have I got, is this actually welcome, what happened the last time. The meeting finishes on schedule. Everyone calls it a good meeting. The one piece of information that would have changed the decision walks out of the room inside somebody’s head.

Nothing in that scene is unkind. Nobody behaves badly. That is exactly the problem – exclusion at work almost never looks like exclusion. It looks like efficiency.

Now run the same meeting again. The senior person speaks last, deliberately, and says so at the outset. The question is not “any concerns?”, which invites silence, but “who here has seen this fail?”. The team leader’s objection gets written down in front of everybody, and a fortnight later they are told what happened to it. That is all. That is what inclusive leadership looks like at work – not a value on a wall, but small, teachable choices about sequence, invitation and follow-up that decide whether what the organisation knows reaches the people deciding.

Over ten years, &Evolve has put the same questions as part of their Cultural Diagnostic, to more than 600 colleagues across six sectors and eight organisations – some 8,800 coded comments. Two challenges surfaced almost everywhere, and most leaders will recognise both. People work in silos. Communication is poor. When those two come back from a survey or a listening exercise, the response is nearly always the same – file them under “communication”, commission a workshop, redraw the org chart and move on.

But silence is not a communication failure. Silence is a rational response to having spoken once and watched it go nowhere. Silos are not a structural accident either; they form where people have concluded that contribution beyond their own boundary is neither wanted nor safe. These are not communication problems. They are what exclusion looks like inside a functioning business, long before anybody uses the word.

Say “inclusive leadership” in most boardrooms and people hear something else entirely: a programme, a budget line, a political argument plenty of organisations are quietly retreating from. That reading has never been more expensive, because the version that governs performance is not about who appears in the photograph. It is a management discipline: teachable, observable, and sitting directly on the P&L.

It is also unusually rare – which is precisely Barney’s point. The advantage is available because so many have vacated it.

One last reason this is urgent, not merely interesting. Think about drag. A body moving through air meets resistance that rises with the square of its velocity: double the speed and you quadruple the drag. Friction that was an irritation at one pace becomes the dominant force at twice the pace. Everything AI promises is speed. Which means every organisation now accelerating is about to discover, precisely and expensively, how much cultural drag it has been carrying.

So the question I would put to any chief executive about to switch on their AI investment is not whether the technology will work. It is this: when it doubles your speed, what happens to your drag?