Tata Consultancy Services (TCS) on 4 November 2026 released the TCS Enterprise AI Study: Intelligence that is Real, a survey of 1,333 senior leaders in 25 countries. It found that 84% rank productivity as AI’s top benefit, yet only 10% let AI agents make decisions without human approval across functions.

The gap between what AI already delivers and how much authority companies hand it is the study’s central finding. Just 9% of executives said their AI agents can securely orchestrate workflows across enterprise applications, a theme Indian IT majors have been building services around to police AI agents.

What did the TCS Enterprise AI Study find?

Security, risk and governance now rank as the leading factor in how and where companies build and deploy AI, ahead of cost, speed and ease of deployment. TCS describes the shift as a Human+AI operating model, in which AI moves from assisting staff to taking on more responsibility where controls allow.

• Productivity first: 84% name productivity as the main source of AI value, followed by customer experience and new products or services.
• Limited autonomy: 10% allow AI agents to act without human sign-off across multiple business areas.
• AI economics under review: 39% are still assessing whether AI projects justify their token and compute costs.
• Open-source traction: 46% report targeted use of open-source AI for selected use cases.
• Fragmented knowledge: 44% say key business knowledge sits with employees or in scattered documents.

“What seems to be holding organisations back, and rightly so, are concerns around security, governance, and resilience,” said Amit Kapur, Chief AI and Services Transformation Officer, TCS, in the TCS press release. He added that 10% of respondents have also seen a revenue uplift from AI.

Model independence and AI sovereignty

Ninety percent of respondents said model independence matters in AI investment decisions, and 34% called AI sovereignty requirements, covering data location, model choice and compliance, a high or very high risk. The figure rises to 43% in Europe. Only 7% have organisation-wide plans to switch AI providers if needed.

Views on jobs are also shifting. The share of leaders who say it is too early to judge AI’s effect on jobs rose from 4% in 2024 to 19% in 2026, according to the study’s key findings.

Why it matters for Indian CIOs

The results echo recent India data. An Avanade report found 75% of Indian firms yet to see AI returns, while a Dun & Bradstreet survey found only 3% of Indian firms have AI-ready data. Governance is also tightening at home, with CERT-In’s new AI security audit guidelines asking for risk tiers, logs and human oversight of AI agents.

For technology leaders, the message is that controls, clean enterprise data and exit plans for AI vendors now decide how far agentic AI can scale.

Key takeaways

• TCS surveyed 1,333 leaders in 25 countries and 12 industries for its 2026 Enterprise AI Study.
• 84% see productivity as AI’s top value, but only 10% let AI agents act without human approval.
• Only 7% have provider-switching plans, even though 90% say model independence matters.

FAQ

What is the TCS Enterprise AI Study?

It is a 2026 global survey by Tata Consultancy Services of 1,333 senior business and technology leaders on how enterprises scale, govern and secure AI.

What is the biggest barrier to autonomous AI agents?

Security, risk and governance ranked as the top factor in AI deployment decisions, ahead of cost, speed and ease of deployment.

Featured image: Tata Consultancy Services office in Madhapur, Hyderabad. Photo: Veera.sj via Wikimedia Commons (public domain)

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter.

"Exciting news! Elets technomedia is now on WhatsApp Channels Subscribe today by clicking the link and stay updated with the latest insights!" Click here!