84% Report AI Productivity Gains, Yet Only 10% Grant Agent Autonomy Amid ROI Struggles

AI Autonomy Remains Limited Despite Widespread Productivity Gains

Enterprises are experiencing significant AI productivity gains, yet remain hesitant to grant systems true autonomy. A TCS study surveying 1,333 senior business and technology leaders across 25 countries found that 84% of executives identify productivity improvements as the leading benefit of AI investment1

1

3

3

. However, only 10% of global enterprises allow AI agents to execute decisions without human approval across multiple business functions1

1

. This reveals a fundamental tension between recognizing AI's operational value and trusting it with decision-making authority.

The gap between confidence and transformation is even more pronounced in specific markets. Indian companies demonstrate greater willingness to delegate to AI-driven processes, with 28% allowing generative AI to complete defined tasks end-to-end without routine human approval, compared to just 19% across all surveyed markets2

2

4

4

. Only 9% of executives globally report having AI agents capable of securely orchestrating workflows across enterprise applications1

1

.

Security Concerns and Governance Challenges Block Full AI Deployment

Security, risk, and governance emerged as the primary factors shaping enterprise AI deployment, ranking ahead of cost, speed, and ease of implementation1

1

3

3

. Amit Kapur, Chief AI and Services Transformation Officer at TCS, explained that security concerns and governance challenges are legitimately holding organizations back. "Deploying AI responsibly, with the right guardrails and enterprise controls, will be the big focus in the next stage of this journey," Kapur stated1

1

.

AI sovereignty requirements present another significant concern. About 34% of respondents identified control over data location, model choice, and regulatory compliance as high or very high organizational risks. In Europe, this figure rose to 43%1

1

3

3

. Despite 90% of respondents considering model independence important in AI investment decisions, only 7% reported having organization-wide continuity plans to switch AI providers1

1

. This suggests concern about vendor lock-in is advancing faster than operational preparedness.

Struggle to Prove Financial Returns Despite Operational Improvements

While companies report operational efficiency improvements, demonstrating financial returns remains elusive. The Avanade report, based on research with 3,600 business decision-makers, found that 75% of Indian respondents said AI has yet to deliver returns that justify the money and effort spent on the technology2

2

4

4

. This compares to 65% across all surveyed markets.

The challenge in proving AI investment ROI is systemic. About 74% of Indian organizations have yet to formally calculate AI returns on investment and have them reviewed by finance teams or senior leadership2

2

4

4

. Globally, 39% of enterprises are still assessing whether their AI initiatives generate sufficient value relative to token and compute costs1

1

. Achal Kataria, Managing Director and CEO of Avanade India, emphasized that "companies that treat AI with the same financial discipline as any other major investment will be best placed to turn early momentum into lasting advantage"4

4

.

Legacy Systems and Integration Barriers Hamper AI Scaling

Integration with existing systems emerged as a critical bottleneck for enterprise AI deployment. Indian respondents reported the highest rate of integration challenges among all surveyed markets, with 48% naming integration with core systems and data as a top challenge, and 41% citing data, cloud, or integration limitations as barriers to becoming an AI leader2

2

4

4

. About 47% of Indian respondents faced integration or scaling problems when deploying AI, the highest of any market surveyed4

4

.

Legacy systems, fragmented data, and integration gaps are preventing organizations from moving AI workloads into production at scale1

1

. About 44% of respondents reported that important business knowledge remains embedded in employees' expertise or fragmented documentation, potentially limiting AI systems' access to enterprise context1

1

3

3

. This data coherence problem undermines AI's ability to operate effectively across enterprise applications.

Confidence Outpaces Transformation in AI Maturity

A striking disconnect exists between business leaders' confidence in their AI capabilities and the actual transformation of operations. The Avanade report revealed that 82% of Indian leaders surveyed have high or complete confidence in their organization's AI use, yet only 23% say their teams have fundamentally reshaped how they operate around it—a gap of 59 percentage points4

4

. Across all surveyed markets, 72% of business leaders are confident in their AI effectiveness, but only 18% report fundamental operational transformation4

4

.

Despite limited measurable impact, 46% of Indian respondents say AI initiatives in their organization are reported as successful4

4

. Ben Beath, Chief AI Officer at Avanade, warned that "too many organizations are measuring AI progress by activity rather than impact. Deploying tools, running pilots and improving productivity are important milestones, but they are not the destination"4

4

.

Regional Variations Show Different AI Priorities

North America demonstrated particular strength in treating AI as a strategic enterprise capability, with 15% of respondents reporting this approach1

1

3

3

. The UK and Ireland stood out for operational resilience, showing comparatively strong levels of AI continuity planning and governance integration3

3

.

India leads in reported operational improvements, ranking first among all surveyed markets for significant, measurable improvement in operational efficiency at 44% compared to 36% across all markets2

2

4

4

. Employee productivity improvements were reported by 42% of Indian respondents against 36% globally2

2

. Indian organizations are also moving past pilots faster, with only 19% making significant investments in proofs of concept, the lowest of any market surveyed4

4

.

Workforce Impact and Open-Source Adoption Trends

Uncertainty about AI's effect on employment is growing. The proportion of leaders saying it is too early to determine AI's impact on jobs rose from 4% in 2024 to 19% in 20261

1

3

3

. This suggests the Human+AI operating model is evolving in unpredictable ways as AI progresses from assisting employees to taking on greater responsibility.

Meanwhile, 46% of organizations report targeted deployment of open-source AI for selected use cases1

1

3

3

. Leadership understanding also faces scrutiny, with 93% of Indian respondents agreeing that leaders could do more to understand AI's capabilities and limitations, and 77% believing senior leaders underestimate how much organizational change is needed to realize AI's full value4

4

.