Why India Inc's AI Rollout Is Outrunning Its Own Workforce

Artificial intelligence is increasingly embedded in enterprise operations, but employees and management structures are struggling to keep pace.

According to IT infrastructure services company Kyndryl's second annual People Readiness Report, AI adoption is accelerating across organisations even as confidence in workforce preparedness declines. The study surveyed 1,100 senior business and technology leaders across eight countries and paints a picture of enterprises investing heavily in AI while falling behind on the organisational changes needed to make those investments pay off.

The report finds that 57% of organisations now have AI embedded in core business processes or deployed broadly across the enterprise, up from 35% a year ago. Yet only 23% of leaders believe their workforce is fully prepared to work with AI, down six percentage points from 2025.

The disconnect is already showing up in business outcomes. Only 32% of organisations have achieved at least one of their top two AI objectives, while just 11% have achieved both. Nearly four in five leaders (79%) believe AI adoption will outpace their workforce readiness, governance frameworks and operating models before their organisations can catch up.

Kyndryl CIO Kim Basile described the findings as a reminder that AI transformation extends well beyond technology deployment. As organisations redesign workflows around AI, they are discovering that people, processes and leadership require far more attention than initially anticipated.

Chief Human Resources Officer Mark Paukel added that the organisations making the greatest progress are those aligning skills, roles and decision-making with changing ways of working instead of simply layering AI onto existing job structures.

Among those organisations is a small group Kyndryl calls “Pacesetters.” They account for just 9% of respondents but consistently outperform their peers by combining workforce redesign, governance and change management into a single transformation programme.

These companies are 1.5 times more likely to report AI-driven revenue growth and 1.6 times more likely to deliver stronger product innovation. They are also twice as likely to have fully implemented every governance practice tracked in the study.

The obvious question is what separates this 9% from everyone else.

The measurement problem

For Neeti Sharma, CEO of TeamLease Digital, the answer begins with how organisations evaluate workforce readiness.

"Most organisations run upskilling workshops but never measure whether skills actually move into daily workflows," she told AIM.

Instead of tracking attendance or course completion, Sharma argues that companies should measure whether employees apply AI to role-specific tasks after training.

Neelabh Shukla, chief business officer at Careernet, believes the problem runs deeper than measurement.

According to him, organisations continue to train employees to operate AI tools without redefining how work itself changes once those tools become part of everyday decision-making.

Most training programmes focus on software usage, he said, but rarely address changes in accountability, collaboration or decision-making. Middle managers, who ultimately determine whether AI becomes part of everyday operations, are frequently left out of transformation programmes despite acting as the bridge between leadership strategy and execution.

What the Pacesetters are doing differently

For Sharma, Kyndryl's Pacesetters stand out not because they spend more on AI but because they execute transformation in the right sequence.

"The 9% Pacesetters pair role redesign with change management and governance simultaneously, so employees are not retrained for jobs that no longer exist," she said.

Drawing on TeamLease Digital's work with global capability centres, Sharma said successful organisations reinforce new behaviours through managers rather than relying on one-time learning programmes.

Shukla approaches the same issue through organisational design.

Companies seeing stronger AI outcomes redesign roles, responsibilities and decision rights alongside technology deployment rather than treating workforce training as a later exercise.

They also connect reskilling directly to live business projects rather than to generic online courses. Equally important, they spend time explaining why the transformation is taking place, reducing the resistance that quietly undermines many AI programmes.

The missing KPI

Both experts agree that enterprises continue to measure AI investments through technology metrics while overlooking workforce readiness.

"It's the absence of a readiness KPI sitting alongside the technology KPI on the same dashboard," Sharma said.

Shukla suggests organisations should instead monitor indicators such as internal mobility into AI-enabled roles, the speed at which employees move from supervised to independent AI use, and whether teams genuinely adopt new workflows rather than reverting to old habits.

Employee trust in AI systems, he added, should also be treated as a leading indicator of long-term adoption.

Hire or reskill?

The report also highlights growing pressure in the AI talent market, with 52% of organisations reporting increased hiring challenges.

For Sharma, that reinforces the case for internal capability building.

The market simply does not have enough AI-collaboration talent to meet demand, making reskilling existing employees the faster and more economical option. Most organisations can build meaningful capability within three to six months instead of competing for a limited external talent pool.

Shukla broadly agrees but believes selective hiring still has a role.

Organisations should recruit a small group of specialists capable of establishing internal expertise before scaling capability through structured reskilling across the wider workforce, particularly in roles requiring judgement about when to trust or override AI-generated recommendations.

Which comes first: redesign or training?

Here, the two experts diverge. Sharma argues against redesigning jobs before employees have the skills to succeed in them.

According to her, changing roles without preparing employees erodes trust because workers feel change is being imposed on them instead of developed with them. She advocates redesign and capability building in parallel, supported by strong change management.

Shukla sees sequencing differently.

"Role redesign has to lead, not follow. Training people for a role that's about to change wastes effort and creates frustration when the goalposts move," he said.

He argues organisations should first identify which roles will be fundamentally reshaped by AI, redefine expectations for those positions, and only then build training around the new responsibilities.

Because AI capabilities continue to evolve rapidly, he believes organisations should treat redesign, training and refinement as a continuous cycle rather than a one-time transformation.

Governance may become the biggest differentiator

Underlying the workforce challenge is a governance gap that grows more significant as AI assumes greater responsibility.

Kyndryl found that only 33% of organisations have clearly defined policies specifying which decisions AI can and cannot make. Just 27% maintain a formal registry to monitor AI systems across the enterprise.

At the same time, 81% of leaders expect AI agents to make consequential business decisions within the next year, even though only 25% say they completely trust AI to operate without human oversight.

The findings suggest governance, rather than technology procurement, could become the defining competitive advantage in the next phase of enterprise AI adoption.

Kyndryl also cited Gartner's estimate that global AI spending will reach $2.52 trillion in 2026, up 44% from the previous year. Yet while technology budgets continue to expand rapidly, investment in workforce readiness remains comparatively modest.

For now, the message from industry leaders is remarkably consistent. AI technology is advancing quickly, but organisational readiness is not. Enterprises that succeed are unlikely to be those deploying the most sophisticated models. They will be the ones who redesign work, establish governance, involve managers and measure workforce readiness with the same rigour they apply to AI itself.