A new global study from Kyndryl has landed a number that should worry any business leader who thinks they’ve ticked the AI box by buying licences and deploying tools: only 23% of organisations say their workforce is actually ready for AI. That figure is down six points from the same study in 2025.
The Kyndryl 2026 People Readiness Report, released on 25 June, surveyed 1,100 senior business and technology leaders across eight countries. The headline finding is a straightforward contradiction: organisations are deploying AI at scale while simultaneously becoming less confident their people can use it well.
The Numbers Don’t Add Up — On Purpose
Here is the tension the report captures. A full 57% of the organisations surveyed have broadly deployed AI or embedded it into core business processes. Another 77% have scaled generative AI across multiple functions. These are not pilot programmes. These are live, production deployments touching real workflows.
And yet:
- Only 23% say their workforce is ready for AI (down from 29% in 2025)
- 79% agree that the speed of AI will outpace their organisation’s workforce, governance, and operating models
- 52% say it has become harder to find employees with the right skills to advance their AI strategy
- 81% expect AI agents to make impactful decisions within the next year, but only 25% completely trust AI systems operating without human oversight
The pattern is consistent: organisations are accelerating adoption while readiness — the human side — lags further behind with each passing year.
Why Readiness Fell While Deployment Grew
This seems counterintuitive. Shouldn’t more AI usage build more AI fluency?
Not when deployment outpaces education. When tools are rolled out without structured training, people learn workarounds rather than fundamentals. They use AI reactively — copy-pasting outputs, clicking suggestions — without understanding how to prompt well, how to evaluate outputs, or how to redesign processes around what the technology can actually do.
The Kyndryl report puts it plainly: AI success is not driven solely by different strategies, use cases, or technologies. It is driven by whether organisations redesign work and manage those changes throughout their people.
Sixty-one percent of survey respondents say their organisations have already redesigned roles to accommodate AI. Twenty-four percent are creating entirely new roles focused on AI management. These are real structural changes — but they’re meaningless if the people stepping into those roles aren’t equipped to handle them.
The Governance Gap Is Just as Real
The trust figures in the report are telling. When 81% of business leaders expect AI agents to be making consequential decisions within the next year, but only 25% trust AI operating without human oversight, you have a structural problem: the expectation is running well ahead of the confidence.
This matters because agentic AI — AI that takes actions, not just provides answers — requires a fundamentally different relationship between employees and technology. People need to understand what the agents are doing, be able to review outputs critically, escalate exceptions, and recognise when the system is operating outside its reliable range.
That is not something you get from a lunch-and-learn about ChatGPT.
What This Means for Business
A few practical implications from where the data sits today:
The deployment bill comes due in people costs. Buying AI tools and deploying them at scale is the easy part. The hard part is the change management: redesigning processes, upskilling teams, establishing governance, and building the institutional knowledge to use these tools well. Organisations that skipped this step in 2024 and 2025 are now paying for it in lower-than-expected ROI and widening confidence gaps.
The skills market is tightening. Fifty-two percent of leaders say finding people with the right AI skills has become more challenging. That means external hiring is not a reliable fix — organisations that don’t build AI capability internally are going to feel the squeeze as the competition for talent intensifies.
Human oversight is not optional — it’s the product. The finding that only 25% of leaders trust AI without human oversight should not be read as a reason to slow down AI adoption. It should be read as the business case for investing in people who understand how to oversee AI well. The value of agentic AI comes from what humans do with it, not despite humans.
Readiness is a competitive advantage. The minority of organisations in the 23% who do feel workforce-ready are operating with a structural advantage. Their people can extract more value from the same tools, catch errors before they compound, and adapt faster as the models and platforms evolve.
For organisations trying to close the gap, the path is deliberate rather than frantic: structured training on the tools people actually use, defined workflows for where AI assists and where humans decide, and a culture that rewards learning over familiarity.
Enterprise DNA’s learning platform exists precisely for this moment. Data professionals who understand how to work with AI — not just use it — will be the ones who determine whether an organisation’s AI investment pays off. Whether your team needs to get serious about Power BI, Python, SQL, or building with AI models directly, the gap the Kyndryl report describes is one that can be closed with the right curriculum and the right commitment.
Source
Kyndryl