Almost everyone is using AI to run the old business faster. Only a third are building a new one.

Your AI programme is working. Costs are down, tasks are faster, teams report real productivity gains, and the dashboard is a wall of green. And yet, if you look honestly, the business is the same shape it was two years ago: the same model, the same offering, the same way of making money, just cheaper to run. That can feel like success. Often it is a trap, the efficiency trap, and new data shows most organisations are walking straight into it while congratulating themselves on the way.

Deloitte's State of AI in the Enterprise 2026 puts numbers on it. Twice as many leaders as last year report transformative impact from AI, which sounds like a breakthrough until you read the next figure: only 34% say they are truly reimagining the business. The rest are applying AI to the existing model. Two-thirds, 66%, report productivity and efficiency gains, which is genuine value but also the most common and least differentiating kind. The rare, valuable move, redesigning what the business actually does and how it does it, is happening in barely a third of companies. Most are automating the old business. Almost no one is building a new one.

What is actually wrong with using AI for efficiency?

Nothing, until it is all you do. Efficiency is real value and worth banking. The problem is that it is the easy, visible, competed-away prize. If AI makes your existing process 30% cheaper, it does roughly the same for every competitor with access to the same tools, and the saving erodes to the customer soon enough. Efficiency keeps you in the game; it does not win it. The advantage that lasts comes from using AI to do something the old model simply could not, a new offering, a new cost structure, a different organisational shape, and that is the 34% move that most programmes never attempt.

Why do so many organisations stop at efficiency?

Because it is the path of least resistance, and every incentive points that way. Automating an existing task is a clean project: a clear owner, a clear before-and-after, and a saving you can bank this quarter. Reimagining the business is none of those things. It is ambiguous, it crosses functions, it threatens existing structures and roles, and its payoff arrives slowly. So the organisation quietly defaults to the tractable version, ships a portfolio of automation projects, and calls it an AI strategy, when what it really has is an efficiency programme wearing a strategy's clothes. The dashboard stays green precisely because nothing fundamental is being risked.

How do I make the reimagining move?

By treating AI as a question about your business model, not just your cost base.

1. Separate "cheaper" from "different." Track both: how much AI is saving on the old model, and what genuinely new capability it is creating. If the second column is empty, you are in the trap.
2. Ask what the old model could not do. The reimagining prize is the offering, service or economics that was impossible before AI. Start there, not at the cost line.
3. Make redesign a leadership project, not an IT one. It crosses functions and threatens structures, so it has to be owned at the top rather than delegated to a rollout team.
4. Do not let efficiency wins masquerade as transformation. A green productivity dashboard can hide the fact that nothing about the business has actually changed.
5. Accept the slower payoff. Reimagination compounds later; efficiency banks now. Fund both, but do not let the quick win crowd out the durable one.

What does this change for me as a leader?

It changes how you read your own AI scorecard. A programme delivering efficiency looks like a win and can still be a strategic dead end if the business underneath it is unchanged. The 34% who are reimagining are not smarter about the technology than everyone else. They are braver about the business: willing to use AI to change what the company does, not merely how cheaply it does the same old thing. Efficiency is the comfortable half of the opportunity, and it is the half that quietly disappears as everyone captures it.

That is the real meaning of the end of business as usual: the winners are not optimising the old model to the last percentage point, they are rebuilding it around what AI now makes possible. Your durable advantage will not come from running yesterday's business more cheaply, because so will everyone else. It comes from what is distinctly yours to build, using AI to do the thing your competitors' efficiency programmes will never reach. Take the saving, but do not mistake it for the strategy.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.