AI returns are up, but readiness is down – Kyndryl on why people hold the key

AI is now paying off for organizations – yet they are less ready for what comes next than they were a year ago.

That is the central tension in Kyndryl’s 2026-27 Readiness Report, which surveyed 3,700 leaders across the globe.

Among organizations investing in AI, 64% now report positive returns – up from 54% in 2025 and 42% in 2024.

By contrast, the average proportion of leaders who say they are completely ready to manage external risks dropped seven points from 2025 to 24%.

“AI is creating value, but AI is raising the bar for readiness,” Kyndryl Chairman and CEO Martin Schroeter stated at a press briefing on the report. “Adopting new technology and being ready to put it to work are definitely not the same thing.”

Readiness is not “destination,” in Schroeter’s words. It needs to be a continuous journey, and Kyndryl research identifies three dimensions leaders need to get right if AI readiness is to keep pace with value.

1. Get your data AI-ready, not just your systems

Seven in 10 leaders describe their IT estate as mostly or fully modernized, yet 62% said they need further modernization before they fully realize AI’s potential. 45% also noted that innovation was being delayed by foundational issues in the tech stack.

The biggest gap is data. 40% said their data is not ready for AI at scale. Plus, those without the right data foundation reported failed AI investments at twice the rate of their peers (35% vs 17%).

To close it, Kyndryl’s report advised organizations work backwards from what AI requires, staring with the data, workflows, and infrastructure – this means tackling data quality, access and integration from the outset.

The other crucial foundational component of AI is skills. Asked what they would change about their AI rollout, leaders’ most common answer was investing more in employee training and AI skills (37%), just ahead of data quality and technology readiness (36%).

Even technology teams are not immune from this skills gap – only 30% said their IT skills and talent are completely AI ready, down seven points from last year.

That’s why AI readiness for Schroeter is more than a technology challenge, it’s a people one: “Bringing people along [the journey] is as important to making AI work as the technology itself.”

Answering UNLEASH’s question in the press briefing, Schroeter shared that if there’s one thing leaders should be acting on it’s “creating those persona-based education and skills opportunity so your teams…can move confidently.”

Kyndryl CIO Kim Basile added in the briefing: “Ultimately, people don’t learn AI by reading about it; they learn…by using it hands-on.”Kyndryl CIO Kim Basile added in the briefing: “Ultimately, people don’t learn AI by reading about it; they learn…by using it hands-on.”

2. Redesign work as one leadership team

Skills are only part of the AI challenge – to bring people on the AI journey organizations also need to move from just applying AI to existing ways of working by reinventing how work gets done.

84% of organizations/leaders are deploying AI in workflows, yet just 27% have redesigned roles around AI and just one third say their workforce is completely ready to use AI.

Reinventing work appears to pay off, according to Kyndryl’s data. Organizations that have both redesigned roles around AI and put formal change management in place report 69% positive AI returns, compared to 53% for the rest.

Getting leadership on the same page is a crucial piece of the work redesign puzzle.

Organizations where leaders report no misalignment on AI see positive returns at 70%, compared with 55% where there is some disagreement at the top. Misalignment is common between the CHRO and CTO or CIO on AI strategy (33%).

Asked by UNLEASH  about how CIOs and CHROs should work together, Basile highlighted the importance of that partnership, while adding “all C-Suite leaders need to really understand how their business can run differently, what the technology can do,” and how to “bring people along on this journey.”

3. Make governance an accelerator, not a brake

Taking people on the AI journey raises a practical question: Who is in charge when AI is doing part of the work?

Basile broke that down into four questions for leaders to ask themselves whenever AI changes a workflow:

• “Who’s making the decision?”
• “When should someone intervene?”
• “Who owns the outcome?”
• “How do we know it's actually working?”

“If those answers aren’t clear, you can have plenty of [AI] adoption without the desired business outcomes.”

Kyndryl’s research found that 40% of leaders say uncertainty about AI governance makes it hard to know where accountability lies.

When asked what they would change about their approach to AI, one third of leaders said they would have implemented stronger governance, risk management and compliance frameworks form the outset.

The data shows that those who take action on AI governance are pulling ahead – they’re more ready to implement AI (46% vs 27%) and see higher positive AI returns (76% vs 63%).

This backs up Basile’s view that “we need to stop treating governance as the opposite of speed.” Kyndryl’s report goes further, urging organizations to build governance to “enable change, not just avoid failure.”

Rather than a final hurdle once an AI solution is built, risk assessment and oversight should be designed into how the technology works, the report noted. The aim is not to remove all uncertainty but to set out how the organization will make decisions and respond when conditions change.

As Schroeter put it: “Enterprise AI is an accumulation of decisions.”As Schroeter put it: “Enterprise AI is an accumulation of decisions.”

For organizations, the lesson across all the three dimensions is that deploying AI and being ready for it are different things – the gap between them is mostly about people, roles, and accountability.