The TCS Enterprise AI Study shows executives overwhelmingly view productivity as AI’s top value, with 84% emphasizing efficiency gains, while only 10% trust AI agents to make autonomous decisions across multiple functions and 9% say AI can securely orchestrate workflows across enterprise applications.

Despite productivity gains, model independence matters to 90% of respondents, and 34% see high or very high data-location and compliance risks related to sovereignty, with Europe particularly concerned (43%). Only 7% report organization-wide continuity plans for provider switching.

AI economics remain unsettled: 39% are evaluating whether the value of AI justifies token and compute costs.

As AI becomes more integrated, success will hinge on guardrails, governance, and accountability to enable trusted, higher-autonomy AI use.

Amit Kapur of TCS says the next adoption phase depends on robust controls and governance to enable responsible, higher-autonomy AI deployment.

Kapur also emphasizes the need for strong guardrails to scale AI responsibly.

Regulatory developments in India and globally complicate cross-border deployment, with calls for governance that retains human oversight for high-stakes decisions and for clearly defined delegated authority and accountability across multiple providers.

Security, risk, and governance emerge as the top factors shaping AI development and deployment, outranking cost, speed, and ease of deployment.

The study flags security, governance, and resilience as the next major focus areas for responsible AI deployment and enterprise controls as use expands.

Experts stress the need for clear contracting and governance safeguards, including liability allocation in vendor agreements, AI-specific indemnities, audit rights, and defined limits of AI authority.

Legal questions about accountability arise as AI agents execute transactions or negotiate terms, including who bears responsibility for errors and whether AI actions can legally bind an enterprise.

A shift toward a Human+AI operating model is underway, with AI taking on more autonomous tasks under appropriate human oversight and safeguards.