In Kyndryl’s 2026 People Readiness Report, there is a clear tension in the once rosy AI adoption story. AI adoption is accelerating, but workforce readiness is moving in the opposite direction. That seems counterintuitive, but in context it makes sense. Corporations remain sold on the benefits. Seventy-seven percent say generative AI has already been scaled across multiple functions, and 57% report that AI is either broadly deployed or embedded in core business processes.
Yet only 23% of leaders say their current workforce is ready to use AI effectively, down six points from last year. The upshot? Organizations have discovered that pushing AI out of proof-of-concept and into production is not just a technology challenge. In fact, the least challenging part may be the technology. Leaders rate IT infrastructure readiness at 35%, but workforce readiness, governance and compliance, and organizational culture trail at 23%, 23%, and 25%, respectively.
This reflects a creeping realization that throwing AI tools at employees and asking them to innovate is more prayer than strategy. Allowing the workforce to determine how best to use AI is essential. But it has to happen inside a more structured process of AI product design, workflow redesign, change management and human-systems thinking. A key learning is that employee readiness is lumpier than many organizations first understood. Support for AI appears to weaken deeper in the organization: executives estimate that 50% of executive leaders are enthusiastically embracing AI, compared with 31% of individual contributors and 30% of entry-level employees.
say generative AI has already been scaled across multiple functions.
say AI is embedded in core business processes or deployed broadly; in 2025, 35% said AI was fully integrated across their organizations.
of leaders say their current workforce is ready to use AI effectively, down six points from last year.
From deployment to absorption: AI shifts left
AI’s first act was about access, experimentation and deployment. AI’s second act is about adoption and absorption: whether people understand how their work is changing, whether roles and workflows have been redesigned, whether governance is in place, and whether employees trust the systems they are being asked to use.
Survey respondents still strongly believe in the AI story and continue to invest. We believe it, too, and have seen core use cases where AI has moved the needle, including cybersecurity, network and systems management, workflow automation, support operations, and complicated forms of pattern matching previously impossible with earlier generation Robotic Process Automation (RPA) approaches. In IT delivery and modernization, we have seen enormous gains driven by agentic progress.
However, the ROI returns remain a work in progress. Only 32% of respondents say they are experiencing one of their top two desired AI outcomes, and just 11% say they are experiencing both. Efficiency gains are showing up first, with 38% reporting improved operational efficiency and productivity. Harder outcomes are less common. Only 14% report additional revenue growth through AI-driven innovation and 11% report innovation in new products, services and business models. This is like the old productivity paradox, the technological progress shows up everywhere but in the profit statement.
What does this all mean? We can easily see places where AI makes a meaningful efficiency or quality contribution. However, the easy part was giving people access to tools. The harder part is redesigning the organization so those tools can create durable value and more tangible outcomes.
No single AI golden path, but four pillars
We are also learning that there is no single “right” way to implement AI at the organizational level. Organizations may optimize for efficiency, augmentation, innovation, customer experience, security, modernization or resilience. Those are strategic choices, not necessarily stages of maturity. One common myth is that companies start with AI for efficiency and then naturally evolve toward augmentation over time. The report complicates that story. Organizations focused on efficiency, and those focused on augmentation can look remarkably similar in how they prepare: redesigning roles, investing in change management, building governance, and developing workforce capability.
Two companies can be equally ready for AI while making very different bets about how AI should shape their workforce. There may be no single golden path for AI strategy. That said, there do appear to be four common pillars upon which the best AI rollouts build: redesign work, manage change, build governance, and prepare people for the new operating model. Note that technology is not among those four.
AI’s second act is not about access to tools. It is about whether organizations can redesign work, prepare people and build the trust needed to turn AI into durable business value.
Closing the trust gap and elevating human judgment
Respondents gave us muddled messages on trust. The report finds that 81% of organizations expect autonomous AI agents to make decisions with material business impact within the next year, and 66% say AI has already been given permission to read and write from core systems of record autonomously, without approval from a person. But only 25% of leaders say they completely trust AI systems to operate without human oversight. And only 24% completely trust AI agents that interact with customers without employee oversight. Fearful of missing out, leaders have leaned into AI for critical work even as they expressed concerns about AI behaving badly. The proper framing should not be whether the agents present business or operational risk but, rather, whether they present acceptable risk and how that risk can be addressed through applying guardrails and compliance and governance processes at the code level.
Yet, even as AI is on the rise, we also heard that human judgment is becoming more essential to helping AI do better — and vice versa. Ninety-five percent of respondents agreed that roles have evolved to require collaboration with AI rather than the replacement of labor. Eighty-two percent say decision-making authority is increasingly shared between people and AI systems. And 24% say they are creating new kinds of roles focused on managing the outputs and workflows of AI.
Readiness is the real AI strategy
The clearest corrective to AI hand-wringing is the report’s Pacesetter group. These organizations have done three things together: redesigned roles around AI, implemented AI-focused change management, and built a workforce they consider ready for AI-related change. Only 9% of organizations qualify. But they are much more likely to report high-value outcomes: 41% report innovation in new products, services and business models, compared with 26% of others; 40% report increased revenue growth due to AI, compared with 27% of others.
The next phase of AI will be defined less by who deploys the most tools and more by who builds the strongest capability in their workforce to leverage AI effectively, with confidence the systemic guardrails put in place will protect against the worst excesses and mistakes. The companies that win with AI will be the ones that focus on people first — redesigning work, preparing people, building trust, and treating workforce readiness as the operating system for AI transformation. Even the most effective, well-crafted tool is useless if the operator is not an expert. So, too, with AI.
