Productivity leads enterprise AI, but autonomy remains scarce
TCS surveyed 1,333 senior leaders in 25 countries. Eighty-four percent cite productivity, but only 10% allow AI agents to execute decisions without human approval.

Why it matters
- TCS's Nov. 18, 2026 survey covered 1,333 senior leaders across 25 countries.
- 84% of executives rank productivity as AI's leading benefit, while 10% report a revenue uplift.
- Only 10% allow AI agents to execute decisions without approval across multiple business areas, and 9% have agents that can securely orchestrate workflows.
- Ireland's no-approval rate is 7%; 39% call sovereignty a high or very high risk, compared with 34% globally.
- Ninety percent value model independence, but only 7% have organisation-wide provider-switching continuity plans.
By Elena Vasquez
84% of executives in a TCS survey rank productivity as AI's leading benefit, while only 10% allow agents to execute decisions without human approval across multiple business areas.
TCS published the global study on enterprise AI adoption on Nov. 18, 2026. The research drew 1,333 responses from senior business and technology leaders across 25 countries, including 44 respondents in Ireland from 11 industries.
Together, TCS's figures describe an autonomy gap: companies report productivity from AI, but few let systems act across several functions. The commercial stakes are direct. TCS says security, governance and resilience matter more as AI gains access to applications, workflows and decision-making.
What do enterprises say AI delivers?
TCS's survey ranks productivity first, followed by improvements in customer experience and the development of new products or services. Only 10% of respondents said they had seen an uplift in revenue, according to Amit Kapur.
TCS's figures show a productivity-first business case, with revenue gains reported by a much smaller share of executives. TCS also found marked regional differences, with North America showing stronger scores on some enterprise AI foundations while Europe recorded greater concern about sovereignty.
Why are enterprises keeping agents under tight control?
Only 9% of global enterprises said they have AI agents that can securely orchestrate workflows across enterprise applications. TCS says many companies remain cautious about moving from tools that assist staff to systems that can act on their own, despite a growing push to embed AI in business processes.
Security, risk and governance emerged as the main factor shaping deployment decisions worldwide, ahead of cost, speed and ease of deployment. TCS links that security focus to businesses giving AI broader access to applications, workflows and decision-making processes.
The two autonomy measures capture different barriers. One concerns authority to decide without human approval. The other concerns secure coordination across enterprise applications. TCS's study therefore places enterprise controls inside the agent design problem, rather than treating them as a later safeguard.
Amit Kapur, Chief AI and Services Transformation Officer at TCS, said:
“We are seeing significant progress in scaling AI for enterprises.
The shift is clearly in AI moving from assisting people to taking more autonomous action with deep access into an enterprise.
Productivity is clearly where enterprises are seeing the greatest value from AI today, with 84% of executives ranking it as the leading source of value from their AI investments, though 10% have also indicated that they have seen an uplift in revenue.
What seems to be holding organisations back - and rightly so - are concerns around security, governance, and resilience. Deploying AI responsibly, with the right guardrails and enterprise controls, will be the big focus in the next stage of this journey.”
How does Ireland compare with the global picture?
TCS's Ireland subset shows a different deployment priority. Irish leaders named cost, speed and ease of deployment as the main driver, while global respondents put security, risk and governance first.
TCS's survey also records less autonomy in Ireland. Seven per cent of Irish enterprises allow AI agents to execute decisions without human approval across multiple functions or business areas, compared with 10% globally.
On model independence, 86% of Irish respondents called it important in AI investment decisions, compared with 90% worldwide. Ireland therefore trails the global figure on willingness to preserve flexibility between models.
Concern about sovereignty moves in the opposite direction. Thirty-nine per cent of Irish respondents identified AI sovereignty requirements as a high or very high organisational risk, above the global figure of 34%. Across Europe, 43% described those requirements as a high or very high organisational risk.
TCS says sovereignty requirements include control over data location, model choice and compliance. Model independence concerns investment flexibility, while sovereignty adds operational and regulatory control over where data sits and which systems companies can use.
TCS also found that the UK and Ireland stood out on operational resilience, with comparatively stronger levels of AI continuity planning and governance integration.
Can enterprises change AI providers?
TCS found that concern about resilience and dependence on particular AI providers was rising faster than contingency planning across the organisations it surveyed. Ninety per cent of respondents said model independence matters in investment decisions.
Only 7% reported having organisation-wide provider-switching continuity plans. That leaves a wide gap between recognising the value of model independence and preparing to replace a provider if prices, service or technical conditions change.
Are the economics convincing enough?
TCS found that 39% of enterprises were still assessing whether AI initiatives generate enough value relative to token and compute costs. The result leaves cost as an active deployment test for many buyers, even when executives report productivity gains.
Another 44% of respondents said important business knowledge remains embedded within employees or scattered across fragmented documentation. TCS says this internal data and knowledge gap may limit an AI system's ability to draw on consistent enterprise context.
The two findings point to separate obstacles. Companies must show that compute spending produces enough value, then provide systems with reliable knowledge from across the organisation. Neither issue disappears merely because agents gain more autonomy.
Where is open-source AI being deployed?
Globally, 46% of respondents said they had targeted deployment of open-source AI for selected use cases, compared with 32% in Ireland, according to TCS's survey. TCS describes open-source adoption as progressing at different rates across markets rather than following one uniform path.
What do leaders know about AI's effect on jobs?
The proportion of leaders who said it was still too early to determine AI's effect on the workforce rose from 4% in 2024 to 19% in 2026. TCS says the increase shows that executives are becoming less certain about AI's employment effects as deployment continues.
TCS's result describes uncertainty over timing and outcomes, rather than a measured employment effect. The rise from 4% to 19% came as companies moved from general interest toward wider use of AI across business processes.
How did TCS gather the data?
The study covered senior leaders at organisations with annual revenues ranging from USD $500 million to more than USD $20 billion, spanning 12 industries.
- 1,333 senior business and technology leaders responded across 25 countries.
- The Ireland count included 44 respondents from 11 industries, while TCS describes the full survey as spanning 12.
- Respondents included C-suite executives, divisional leaders, Senior Vice Presidents and Vice Presidents.
Those details place the findings among large enterprises rather than small businesses or individual workers. They also explain why governance, continuity and technical deployment recur throughout TCS's results.
What comes next?
Kapur expects the next stage of enterprise AI to center on responsible deployment with guardrails and enterprise controls. The immediate test for buyers is whether agent autonomy can expand without turning security, governance and resilience into afterthoughts.
For employees, the 19% who called the workforce impact too early to determine in 2026 vastly outnumber the 4% who said so in 2024. The practical measure of progress will be wider authority paired with measurable controls and returns, not productivity claims alone.
TCS's forecast is that agent access will keep widening, while guardrails and enterprise controls become the decisive condition for deployment.
Source: itbrief.com.au
More from Elena Vasquez
Show full bio
Market editor covering technology and product news in artificial intelligence and machine learning at AI In Context.
708 articles
Related articles
- Only 10% of Enterprises Let AI Agents Decide Without Human Approval
- 51% of AI projects miss targets, yet 85% of firms raise budgets
- The EU's Next Chapter in AI Is Agentic, New Analysis Claims
- OpenAI reports frontier AI firms now generate 8.3x more output per user
- Enterprises Kill 46% of AI Pilots Before Production, S&P Global Finds