AI Adoption in India: Turtle Momentum

At a mid-sized auto-components firm outside Pune, the HR head announced at a packed town hall in early 2025 that the company was "going AI-first" in hiring. A resume-screening tool would cut time-to-hire in half. Eighteen months on, the tool still runs - but almost every shortlist it produces is quietly re-checked in the same Excel sheet recruiters used before. Nobody switched it off. Nobody fully trusted it either. On paper, the company had adopted AI. In practice, it had added a step.

That gap - between adoption on a dashboard and transformation on the ground - is the story of AI in Indian workplaces today.

India’s AI adoption numbers are genuinely strong. NASSCOM’s AI Adoption Index found 87% of Indian enterprises were actively using AI as of late 2025, and Deloitte’s 2026 State of AI in the Enterprise report found Indian firms leading global peers on at-scale deployment - 62% in product development, 56% in strategy and operations - with 40% of respondents reporting significant or full AI usage against a 28% global average. EY’s 2025 Work Reimagined survey gave India the highest "AI Advantage" score worldwide.

This same gap between adoption and trust shows up across sectors and industries. Netskope’s India 2026 threat report found that even as organisation-managed AI tool usage rose from 30% to 77% of employees over the past year, personal, unauthorised AI use persisted at scale rather than disappearing (Netskope, 2026) - the digital equivalent of recruiters quietly keeping their old Excel sheet running alongside the new tool. The result, as MIT’s 2025 GenAI Divide study found globally, is that most enterprise AI pilots show no measurable business impact - not because the models fail, but because the workflows and trust around them are never redesigned to match (Challapally et al., 2025).

Yet the same reports carry a quieter thread. NASSCOM notes fewer than 15% of companies have aligned AI strategy with core business strategy, hamstrung by legacy systems and siloed data. Marsh India’s 2026 People Risk report found 43% of HR and risk leaders flagging employee resistance from fear of job loss or change fatigue, and 49% admitting their organisations invest in AI without adequate training. LinkedIn’s January 2026 India survey found 84% of professionals feeling unprepared for an AI-altered job market, despite 87% saying they use AI comfortably. India races ahead on pilots and press releases; it moves like a tortoise on the harder work of trust and role redesign.

Three forces slow the shift from experimentation to transformation. First, a skills mismatch: the Mercer-Mettl India Graduate Skill Index 2025 puts overall graduate employability at 42.6%, and NASSCOM-McKinsey research projects a shortfall of 1.4 million AI-skilled professionals by 2027. Second, legacy infrastructure and data silos make many "AI-first" initiatives cosmetic rather than structural. Third, and least discussed, is the middle-management bottleneck: managers are expected to champion tools that quietly threaten the activities - approvals, headcount calls, quality checks - on which their own performance is judged. McKinsey research finds only about 30% of AI pilots globally ever scale; Indian managers add the burden of translating top-down mandates to teams already anxious about their place in it.

An I-O Psychology Lens

I-O psychology offers a clearer lens than "resistance to change." Bandura’s self-efficacy theory explains why employees who don’t yet trust their own AI competence quietly work around a tool rather than with it - as the Pune recruiters did. The Technology Acceptance Model predicts perceived usefulness collapses fast when a tool isn’t seen as genuinely easier or fairer, which is exactly what Indian data on AI anxiety and "trust gaps" is now picking up. Status-quo bias, well documented in organisational behaviour research, means people default to familiar routines under uncertainty - even when a new tool objectively saves time. None of this is irrational; it is a predictable human response to change introduced without psychological safety.

What This Means, and What to Do

1. Organisations should treat AI rollout as change management, not IT deployment - with role-level clarity on what changes, and safe channels for employees to flag AI errors without penalty.
2. Managers need a seat in pilot design early, and should be measured on adoption and team capability, not legacy activity metrics, so they have a real stake in making AI work.
3. Employees benefit from treating AI literacy as a continuous, low-stakes skill - in the spirit of India’s World Economic Forum Skills Accelerator (Ministry of Skill Development and Entrepreneurship [MSDE], 2026) commitments - rather than a one-time threat.
4. Policymakers and educators, given the employability gap in the Mercer-Mettl data, have a role in embedding applied AI fluency into graduate curricula before students enter workplaces that already assume it.

India is not short of AI ambition, investment, or usage. What it is short of is the delayed work of building trust, skill, and psychological readiness alongside the technology - the work that decides whether a resume-screening tool ends up trusted, or just tolerated. Closing that gap, not the next product launch, is what will turn India’s AI story from turtle-paced experimentation into real transformation.