AI projects underdeliver as firms keep boosting spending

Genpact has published research suggesting many corporate AI projects are failing to deliver the expected business value, even as most companies plan to increase AI spending.

The study covered more than 2,000 enterprise executives across 17 industries. It found that 79% of respondents have AI agents in production, but only 15% said they had fully linked AI investment to profit-and-loss outcomes.

On average, respondents said just 49% of AI initiatives funded in the past 24 months had delivered the expected value. The findings highlight a gap between putting AI systems into use and changing operating models so they affect financial performance.

AI now leads or fully runs 40% of decisions, more than double the level of a year earlier, according to Genpact. Even so, humans still make or audit decisions in at least 87% of cases at every stage, suggesting companies continue to rely heavily on staff oversight.

The report also found that adoption has not generally led to immediate workforce cuts. Among respondents with AI agents in production, 92% said they had either maintained or increased headcount.

Operating gap

Only 20% of those surveyed said they were using AI agents broadly and consistently across functions, while just 9% said they had completed significant operating model transformation. That suggests many deployments remain limited in organisational reach.

The research argues that the main barrier to value is operational rather than purely technical. Many companies, it says, have adopted AI tools faster than they have redesigned workflows, management structures and accountability around them.

Genpact identified a subset of respondents it called AI Leaders, representing 17% of enterprises in its AI Maturity Index. Those companies were more likely to overhaul work across the business, define new roles for employees working alongside AI systems, strengthen data foundations, and tie AI gains directly to named owners and P&L lines.

They were also more likely to govern AI agents across their full lifecycle rather than only at the point of deployment. That includes overseeing how agents act across systems, not just how they retrieve information.

Spending persists

The findings come as companies continue to support AI budgets despite mixed returns. Genpact's survey found that 51% of AI initiatives failed to deliver the expected business value, yet 85% of companies planned to increase AI spending.

That suggests many executives still see AI as strategically important even when current projects fall short. The data also indicates that boards and senior management continue to fund deployments while working through practical questions about governance, integration and measurement.

The survey was conducted among senior executives at companies with annual revenues ranging from USD $1 billion to more than USD $50 billion, spanning 14 business functions. Genpact said the research was supplemented by in-depth interviews with enterprise leaders.

Sanjeev Vohra, Chief Technology and Innovation Officer at Genpact, said the challenge had moved beyond deployment.

"Enterprises have become effective at deploying AI agents. Running them is the hard part," he said. "Agentic Operations connects AI with context-rich process intelligence, human judgment, and governance to capture the autonomy dividend."