top of page

The State of AI at Work: Adoption Without Transformation

11 minutes ago
3 min read

The story most organizations tell themselves about artificial intelligence is a story of rapid progress. Tools have been purchased, pilots have been launched, and nearly every leadership team can now point to some initiative that demonstrates the company is taking the technology seriously. Beneath that reassuring surface, however, the most rigorous data reveals a far less comfortable picture, one that should reframe how senior leaders think about where they actually stand. The gap that matters is no longer between the companies that have adopted AI versus those that have not. It is between the companies that have merely adopted it and the very few that have genuinely transformed because of it.


McKinsey's State of AI 2025 survey, drawn from nearly two thousand respondents across more than a hundred countries, found that 88 percent of organizations now report using AI in at least one business function. That figure sounds like a triumph, however, the same report shows that nearly two-thirds of those organizations remain stuck in experimentation or pilot mode, with only about a third having scaled the technology across the enterprise in any meaningful way. Most telling of all, just 6 percent qualify as high performers, meaning they attribute at least 5 percent of their earnings before interest and taxes directly to their use of AI. The overwhelming majority of companies have adopted the vocabulary of transformation while capturing almost none of its value.


Why Adoption and Value Have Come Apart

The organizations capturing real value are the ones willing to redesign how work is actually done, rather than the ones that have simply accumulated the most tools. McKinsey found that high performers were 2.8 times more likely than their peers to have fundamentally reworked their workflows when deploying AI rather than layering the technology on top of processes that were designed for a pre-AI world. This is the crux of the matter, because a tool dropped into an unchanged process produces a faster version of the old inefficiency. 


This distinction places the burden squarely on leadership rather than on technology. Buying access to a capable model requires a budget and a signature. Redesigning a function around it requires a leader willing to challenge established roles, retrain people, tolerate a period of disruption, and hold a clear point of view about what the organization is trying to become. The reason so few companies have crossed from adoption into value is that the second act demands exactly the kind of sustained, courageous leadership that is always in short supply, and that no software license can supply on its behalf.


What This Means for the Months Ahead

As leaders build their plans for the coming year, the question is no longer whether the organization is using AI, because almost everyone is now. The more revealing question is whether the organization has changed in any fundamental way as a result, or whether it has simply added a layer of impressive tools to a structure that operates much as it did before. The companies that will separate themselves over the next several quarters are the ones whose leaders treat AI as an occasion to rethink the work rather than as a box to be checked. The competitive distance between the two groups is already widening, and the data suggests it will be very difficult to close later for those who mistake adoption for transformation now.


Two coworkers discuss AI code on dual monitors in a modern office, one pointing at the screen while the other listens attentively.

 
 
 

Comments


bottom of page