Despite growing productivity gains from artificial intelligence (AI), only 10% of global enterprises allow AI agents to execute decisions without human approval across multiple business functions, according to a new global study by Tata Consultancy Services (TCS).
The TCS Enterprise AI Study: Intelligence that is Real found that 84% of surveyed executives identified productivity improvement as the leading source of value from AI investments. However, only 9% reported having AI agents capable of securely orchestrating workflows across enterprise applications.
The findings highlight the gap between businesses benefiting from AI-assisted tasks and their readiness to give AI systems greater autonomy in making decisions and executing complex operations.
Security, risk, and governance emerged as the leading considerations influencing how organizations develop and deploy AI systems, ranking ahead of cost, speed, and ease of implementation. TCS said enterprises will need stronger safeguards and controls as AI agents gain access to business applications, sensitive information, and decision-making processes.
The study also found that 90% of executives consider model independence important when making AI investment decisions. However, only 7% of global respondents reported having organization-wide continuity plans that would allow them to switch AI providers without disrupting operations.
AI sovereignty is another growing concern, with 34% of executives identifying requirements involving data location, model selection, and regulatory compliance as a high or very high organizational risk. The figure rises to 43% among European respondents.
Other findings reveal that 39% of enterprises are still evaluating whether AI initiatives deliver sufficient value relative to token and computing costs. Meanwhile, 44% reported that important business knowledge remains embedded in employees or fragmented documentation, potentially limiting AI systems’ access to organizational information.
Open-source AI adoption is also gaining momentum, with 46% of respondents reporting targeted deployments for selected business applications.
The study further highlighted uncertainty surrounding AI’s workforce impact. The proportion of executives who said it was too early to determine AI’s effect on employment increased from 4% in 2024 to 19% in 2026.
TCS said enterprises are moving toward a Human+AI operating model in which AI systems progress from assisting employees to handling more complex tasks with increasing autonomy. However, stronger security, governance, and operational resilience will be essential before organizations can safely expand AI decision-making capabilities.
The study surveyed 1,333 senior business and technology leaders across 25 countries and 12 industries, representing organizations with annual revenues ranging from US$500 million to more than US$20 billion.
The full report, TCS Enterprise AI Study: Intelligence that is Real, is available through the TCS website.
