India loves AI. So why are so many workplace AI pilots failing?

India may be among the world's fastest adopters of workplace AI, but its AI journey is exposing an uncomfortable reality: many enterprise AI pilots are failing to deliver.

According to Salesforce's Agentic Workplace Study, 38% of workers in India say their organisation experienced an unsuccessful AI pilot in the past year — the highest among all markets surveyed and significantly above the global average of 28%.

The disappointing pilots are emerging in one of AI's most enthusiastic markets. The study found that workers in India are 45% more likely than the global average to say AI is now part of their core workflow, the highest across all surveyed countries. AI has clearly moved beyond experimentation and into everyday work. But wider adoption is also creating more discerning users.

The report suggests that Indian employees are not rejecting AI itself, they are rejecting generic AI.

Lacking business context

The biggest reason workplace AI initiatives fail, according to employees, is that the technology often lacks sufficient understanding of business context. Among respondents whose AI tools or pilots did not work out, 34% said the AI failed because its outputs lacked business context, well above the global average of 22%.

Workers also pointed to several other shortcomings. Nearly 31% said AI could not be adequately personalised to their specific role or tasks, while an equal proportion felt the outputs remained too generic to be genuinely useful. Others cited insufficient training (29%) and the growing complexity of managing too many AI tools (28%) within the workplace.

The study also reveals a striking contradiction. Despite being among the world's most active AI users, 49% of Indian workers still describe themselves as AI skeptics, compared with 37% globally. Rather than signalling resistance to the technology, Salesforce suggests the skepticism reflects rising expectations. As AI becomes embedded in everyday workflows, employees expect it to deliver relevant, accurate and contextual support instead of generic responses.

Building trust

The next battle isn't getting employees to use AI, it's making AI useful enough to earn their trust. The report argues that the next phase of AI adoption will depend less on launching more pilots and more on building AI that employees trust. That means connecting AI to enterprise data, tailoring it to specific roles, integrating it into workflows and training employees to use it effectively.

“India's AI story is now about execution. The workforce is experiencing AI in the flow of work, and that experience is creating a more mature expectation from technology — AI that understands their business, supports their role, and helps them make decisions with confidence.

Leaders must now focus on operationalising AI — connecting data, redesigning workflows, building trust, and equipping people to work alongside agents. If this is done well, AI will not just lift productivity. It will define how India competes, innovates, and grows.”

The Salesforce's Agentic Workplace Study was conducted in partnership with YouGov between December 2025 and January 2026. The double-blind online survey covered more than 1,500 desk workers across Australia, India (120 respondents), Japan, Singapore, France, Germany, Italy, the Netherlands, Spain, the United Kingdom, Mexico, the United States and Canada. Respondents represented a cross-section of job roles, industries and organisation sizes, and all had at least a basic familiarity with AI.