What the 21% Do Differently

Eighty-eight percent of organizations are now using AI in some part of their operations. If you had told anyone that number five years ago, they would have called it a revolution. And it is — just not the kind anyone expected.

Because here is the other number: eighty-one percent of those same organizations report no significant bottom-line impact from their AI initiatives. They are using the tools. The results are not there.

That gap — between adoption and value — is the most important story in business right now. And almost no one is telling it correctly.

The finding that changes everything

In February 2026, McKinsey published The State of Organizations 2026, a survey of ten thousand senior executives across fifteen countries. The question they were trying to answer was simple: what separates the organizations getting real, measurable value from AI from the ones that are not?

The answer was not budget. It was not technology choice. It was not sector, size, or geography. The single most correlated factor with meaningful AI success was this: workflow redesign.

Only 21% of organizations have fundamentally redesigned their workflows around AI. Those that have are 2.8 times more likely to report significant EBIT impact from their AI investments. The other 79% are layering AI on top of existing processes and wondering why nothing changes.

Adding AI to a broken workflow does not fix the workflow. It accelerates it.

What workflow redesign actually means

Workflow redesign is not a technology project. It is a thinking project. It requires looking at how work actually happens — the sequence, the handoffs, the decisions, the bottlenecks — and asking: if we were designing this from scratch today, knowing what AI can do, would we do any of this the same way?

Most organizations do not ask that question. They find a task that feels automatable, add an AI tool to it, and measure whether that task now takes less time. Sometimes it does. But the broader system stays the same, which means the broader system continues to produce the same results, slightly faster.

The 21% ask a different question. Not: where can we insert AI? But: what is this process actually trying to achieve, and what is the best possible way to achieve it given the tools now available? That question often leads somewhere surprising. It reveals that entire stages of a workflow exist only because of limitations that no longer apply. It surfaces assumptions that were never examined because they seemed fixed.

The investment ratio nobody talks about

The same McKinsey research surfaced another finding that reframes the entire AI investment conversation. For every dollar an organization spends on AI technology, it should be spending five on people. Not five on training to use specific tools — five on developing the thinking capacity, judgment, and organizational capability to use AI well.

Organizations that get this ratio right are four times more likely to sustain top-tier financial performance over the next decade. The ones that do not — the ones investing heavily in technology while underinvesting in people — are producing the statistics that fill the bottom of that 81%.

What this means for individuals

The organizational finding has a direct parallel at the individual level. The professionals getting the most from AI are not the ones using the most tools, or using them most frequently. They are the ones who have redesigned how they think and work around AI — who have asked, honestly, what changes about how I do this now that AI exists?

That question is harder than it sounds. It requires examining habits and workflows that have become invisible through repetition. It requires a tolerance for disruption during the transition period, when the old way is being dismantled and the new way is not yet smooth. Most people skip this phase. They add AI to how they already work. Then they wonder why they are not getting the results they expected.

The 21% do not skip it. That is the difference. Not talent. Not access. Not luck. A willingness to examine the thing itself — not just add a layer on top of it.