TCS survey: AI boosts productivity, not revenue yet

That matters because AI is moving from pilot projects into operational workflows, yet the benefits are not evenly translating into bottom-line results. While companies are using AI to speed up tasks and augment workers, the study shows only 10% of enterprises allow AI agents to execute decisions without human approval across multiple functions. That indicates a still-cautious approach to automation even as vendors push more autonomous systems.

The restraint is not just cultural. Security, risk and governance now rank ahead of cost, speed and ease of deployment in determining how AI systems are built and rolled out, according to the study. For investors, that suggests adoption will likely remain uneven and capital-intensive, favoring firms with strong enterprise software, consulting and cloud positions while slowing the pace at which AI generates operating leverage for end users.

The survey also points to a second-order bottleneck: 39% of enterprises are still weighing whether AI returns justify token and compute costs. That is important for profitability assumptions across the AI stack, from cloud infrastructure providers to software companies and service integrators that are spending heavily to embed AI in their offerings. Open-source AI is being used selectively by 46% of respondents, underscoring how firms are trying to balance flexibility, cost control and risk.

For the labor market, the message is more ambiguous. Companies are shifting toward a “Human+AI” operating model, but leaders’ uncertainty over jobs is rising, with the share saying it is still too early to determine AI’s effect on employment climbing to 19% from 4% in 2024, according to the study. That suggests AI is becoming a management tool before it becomes a clear employment disruptor, even as firms continue to test where automation can replace or augment human decision-making.

The report also highlights a strategic issue for global investors: AI sovereignty. Ninety percent of respondents say model independence matters in AI investment decisions, while 34% see sovereignty requirements as a high or very high risk, rising to 43% in Europe. That points to a more fragmented market ahead, where regulations, data control and national-security concerns could shape procurement and delay some deployments.

In market terms, the study reinforces the case that AI remains a durable corporate spending priority, but one whose payoff is arriving first in productivity rather than revenue. The bull case is that efficiency gains expand margins and justify continued investment; the bear case is that governance, compute costs and fragmented enterprise data keep monetization slower than the hype cycle suggests. For now, investors should expect AI spending to stay high, but realized returns to be selective.