HCLTech’s latest global research report, The Blueprint for AI Leadership, highlights a paradox: most organizations believe they are ready for AI-driven workforce change, but far fewer have built the systems to make that readiness real. While 79% say they are very or somewhat prepared to train and upskill employees for operational changes driven by Agentic AI, only 33% have a comprehensive organization-wide retraining or upskilling strategy. Nearly two-thirds say their plans are defined but limited in scope.
A similar gap appears in recruitment, onboarding and human-agent orchestration. Organizations are confident in adapting recruitment and onboarding (79%) and blending work across humans and agents (80%). Yet confidence weakens around role redesign, redeployment and consistent adoption at scale.
“Confidence without strategy is a form of organizational wishful thinking. It's comfortable in the short term but creates significant risk as AI capabilities mature and workforce requirements evolve,” says Lester Lam, EVP and Global Head, Advisory and Consulting, Digital Business Services, HCLTech.
Only 27% of organizations are very confident in managing AI-driven change today. AI Leaders (18% of the sample) remain far ahead of AI Followers (60% of the sample), though many expect the gap to narrow within three years. That optimism raises a key question: are organizations measuring AI returns against current execution capability, or against an assumed future state?
The difference is stark. Some 93% of Leaders have a comprehensive retraining or upskilling strategy, compared with only 20% of Followers.
“AI Leaders are treating upskilling as a strategic priority with commensurate investment,” says Lam. “They've made substantial commitments of time, money and executive attention to ensure dedicated learning time is built into work schedules, not just optional evening courses.”
Followers often treat workforce development as a periodic HR initiative rather than embedding it into how work is organized. Only 24% are prepared to manage continuous AI-driven organizational change, versus 52% of Leaders.
“Upskilling efforts that are disconnected from how performance is evaluated or how work is organized will always underperform, because employees correctly read that the capability being developed is not actually required to succeed in their current role,” says Sebastian Reiche, Professor of Managing People in Organizations, IESE Business School.
Followers are also three times less likely than Leaders to encourage experimentation with AI tools and new ways of working. Organizations that embed capability more effectively design work so learning happens naturally, through role rotation, peer feedback and employee agency. They also align work design, performance management and incentives. AI literacy, then, is not just a technical training challenge; it is a work design challenge. Organizations must define what each role becomes when augmented by AI, then build learning around that reality.
Closing the capability gap
For AI Followers, the window to close the gap is narrowing. Leaders are already as far ahead on skills as they are on technology because they have built the workforce infrastructure needed to convert AI investment into compounding advantage: development systems, experimentation cultures and role redesign capabilities.
Followers may also be diagnosing the wrong problem. AI adoption is often framed as employee resistance, but the research suggests otherwise. The leading barrier is skills shortage, cited by 47% of organizations. Employee resistance is cited by only 20%, while leadership misalignment ranks higher at 29%.
Talent dynamics shift with maturity. Attracting and retaining AI-proficient talent is a top challenge for 38% of Leaders, but only 26% of Followers. Mature organizations are already competing for specialized skills, while Followers are still building baseline readiness. Insufficient training programs are cited by 30% of Followers versus 15% of Leaders, reinforcing that the issue is not resistant workers but underdeveloped learning infrastructure.
“AI's transformational potential isn't realized when you implement AI systems. It's realized when people throughout your organizations know how to leverage those systems effectively, can identify new opportunities for AI application and are comfortable working in AI-augmented ways,” says Lam. “The technology is necessary but not sufficient.”
Reiche reinforces the point: “Employee resistance as a narrative is indeed convenient, but it is incomplete.” Workers are often more willing to adapt than assumed, especially when they help shape how tools enter their roles and can see how their work will evolve.
The deeper challenge is structural. The dominant work design model, in which tasks are centrally defined and assigned, is poorly suited to the dynamic reconfiguration AI requires.
“That model produces clarity and control, but it is very poorly suited,” explains Reiche.
When AI enters the workflow, organizations must rethink what roles are for, how work moves between people and systems and where human judgment adds the most value. These are work design questions, yet many organizations lack the frameworks and habits to address them.
Upskilling for the autonomous era
The urgency is increasing. Less than 10% of today’s enterprise AI capability is autonomous, but that share is expected to rise sharply over the next three years across technology, infrastructure, data and analytics. Leaders and Followers remain far apart in preparing for this shift, particularly in managing workforce displacement (38% of Leaders are very prepared versus 19% of Followers) and redesigning roles (51% versus 29%).
Governance, oversight and the ability to challenge autonomous systems will become core workforce capabilities. They will not emerge automatically. Even today, readiness for AI governance is just 43%, well behind readiness to deploy or maintain AI systems.
“Organizations can’t govern what they do not understand,” says Reiche. “Building a workforce capable of meaningful AI oversight requires, first, that people working alongside autonomous systems have enough understanding of how those systems operate to know when something is wrong.”
If AI systems are treated as black boxes and employees simply execute outputs, governance will remain weak. But when work is designed so employees interrogate outputs, flag anomalies and refine systems, oversight becomes part of daily operations. Organizations will also need roles that sit between AI systems and human context, requiring contextual judgment, ethical reasoning and the confidence to challenge system outputs.
Without these workforce systems, AI initiatives may remain promising but limited. Where learning, oversight and adaptation are built into the operating model, organizations are far more likely to turn AI investment into sustained enterprise impact.
