Kyndryl just released its 2026 People Readiness Report. The headline number: only 23% of organizations say their workforce is fully ready for AI. That is down from 29% last year.
In the same period, AI adoption went from 35% of organizations saying AI was integrated across their business to 57%. More than half of enterprises now run AI in core processes. And the workforce that operates those processes is less prepared than it was twelve months ago.
That is not a paradox. That is the natural result of treating AI as a technology deployment instead of an operations redesign.
The Numbers That Matter
The Kyndryl study surveyed 1,100 senior business and technology leaders in eight countries. The findings tell a consistent story: AI is moving faster than the organizations using it.
79% of leaders agree that the speed of AI will outpace their workforce, governance, and operating models. Only 32% have achieved at least one of their top two AI objectives. Just 11% have hit both.
52% say it has become harder to find employees with the right skills for their AI strategy. And only a third have fully stood up change management programs to guide the transition.
These are not technology problems. The models work. The platforms are available. The budgets are approved. Global AI spending is forecast to hit $2.52 trillion in 2026, a 44% increase from last year, according to Gartner. The money is flowing. The results are not.
The 9% Who Figured It Out
The study identifies a group it calls Pacesetters. They represent 9% of organizations surveyed. What separates them is not their AI budget or their model selection. It is three operational decisions:
- They redesigned roles around AI, not just attached AI to existing roles.
- They implemented change management so the workforce understood the new operating model and had guardrails.
- They built workforce readiness before scaling deployment.
The results speak for themselves. Pacesetters are 1.5 times more likely to achieve AI-related revenue growth and 1.6 times more likely to report better innovation in products and services. They are also roughly twice as likely to have fully implemented every governance dimension measured.
None of that is about technology. All of it is about how work is organized.
Why Readiness Dropped While Adoption Surged
The readiness number going down while adoption goes up tells you everything about where most organizations are. They shipped the AI. They did not ship the operating model that makes it useful.
61% of organizations say they have already redesigned roles. But less than a third have fully established change management programs. That gap is where value goes to die. You can restructure a role on paper, but if nobody explained to the person in that role what changed, why it changed, and how to work differently, you have not redesigned anything. You have just created confusion with a new job description attached.
This is the pattern I see over and over in practice. The AI works. The team does not know what to do with it. Ownership is unclear. The workflow around the AI has not been rethought. And then six months later, someone in leadership asks why the ROI is not showing up.
The answer is almost always the same: nobody changed the work. They just added a tool.
The Trust Problem Is an Operations Problem Too
81% of organizations expect autonomous AI agents to make decisions with material business impact within the next year. But only 13% completely trust AI systems operating without human oversight.
That gap is going to produce one of two outcomes. Either organizations slow down agent deployment until trust catches up, which means falling further behind the Pacesetters. Or they deploy agents without trust, which means governance gaps, compliance exposure, and the kind of failures that make headlines.
The organizations with stronger governance report higher workforce trust. And higher trust correlates with better AI outcomes. The report makes this connection directly: trust is not a feeling. It is the output of clear policies, registries, monitoring, and communication about what AI can and cannot decide.
In other words, trust is built through operations. Not through reassurance.
What This Means for You
If you are leading a team or an organization through AI adoption, the Kyndryl data points to one thing: the constraint is not the technology. It is whether you have done the harder work of redesigning how your people operate alongside it.
The 9% pulling ahead are not smarter. They are not spending more. They are doing three things the other 91% are skipping or doing halfway. Redesigning roles. Running change management. Building readiness before they scale.
The readiness number dropped six points in a single year. That means the gap between organizations that prepared their workforce and those that did not is widening right now. Not in the future. Not when agents arrive. Right now.
Every month you wait to do the operations work is a month the Pacesetters compound their advantage. They are not waiting for you to catch up. They are building the muscle memory and the institutional knowledge that makes every next AI deployment faster and more effective than the last.
The model you picked is not the problem. The workflow you did not redesign is.