Tata Consultancy Services is building an enterprise AI stack with a governance layer designed to coordinate AI agents developed by TCS, its clients and third-party vendors. The layer is intended to act as a central control plane across business software—including platforms from SAP, ServiceNow and Oracle—so companies can manage deployments, security and costs as they expand their use of autonomous systems.
TCS executive director and chief operating officer Aarthi Subramanian described the approach in an interview published by The Financial Express on October 12, 2026. The company has not, in that report, announced a named standalone product or a public launch date for the governance layer.
Why enterprise AI needs a control plane
Many businesses are moving from employees using individual chatbots to AI agents that can carry out steps inside business workflows. One agent might retrieve information from a customer relationship system, another could prepare a finance report, and a third could open or update a service ticket. These agents may come from different vendors and operate across systems that were never designed to coordinate autonomous software.
That creates a management problem. Without a shared control layer, IT teams can struggle to see which agents are active, what data they can access, how much they cost, and whether their actions follow company policy. A central governance layer can provide a common place to orchestrate agents and monitor their use, rather than forcing each business unit to manage every integration separately.
TCS says its proposed layer is intended to coordinate agents built internally as well as those supplied by clients and outside technology providers. In practical terms, enterprises will also need clear identity and access controls, audit trails, budget limits, and rules for when an agent must hand a decision back to a person. Those are implementation requirements to test—not capabilities that TCS has publicly detailed for this layer.
The timing: TCS AI revenue crosses $3 billion
The governance push comes as AI-related work becomes a more material part of TCS’s business. In its September-quarter results announced on October 8, TCS reported annualised AI revenue of US$3.1 billion, more than 10% of overall revenue. That annualised figure is a run rate based on the quarter, not revenue booked from AI in that quarter alone. It was up from US$2.6 billion in the June quarter.
TCS said demand is coming from AI-native solutions, AI-led transformation of enterprise systems and more autonomous business services. Banking and financial services, manufacturing, and life sciences and healthcare were among the areas contributing to the momentum, according to The Financial Express report.
However, the AI growth sits alongside a more cautious market for traditional IT services. Reuters reported on October 8 that TCS’s sequential revenue growth was just 0.5% in constant-currency terms—the weakest September-quarter performance in three years—while total contract value for the quarter was US$9.6 billion. TCS management also said clients continued to scrutinise discretionary programmes that lack near-term value.
That makes governance more than a safety feature. If companies are to move AI agents into everyday operations, they need to know whether the systems are reliable, secure and economical enough to deliver measurable results.
What it could mean for Indian businesses
For Indian banks, manufacturers, retailers and healthcare providers, the challenge is not simply choosing a capable AI model. It is fitting agents into existing enterprise software while keeping sensitive data, permissions and accountability under control. A common governance layer could make it easier to oversee agents from several providers, particularly where workflows cross departments or involve regulated information.
TCS has already positioned governance as part of its broader enterprise AI strategy. In April, it expanded its Google Cloud partnership around AI-native operating models, with an emphasis on deploying and managing agentic systems without adding operational risk or complexity. The newly described control layer fits that direction, although the company has not yet published enough technical detail to judge how it will integrate with competing agent platforms.
For customers, interoperability will matter as much as the promise of automation. A useful control layer should work across systems rather than lock an organisation into a single model or vendor. Buyers should also look for evidence of how access is granted and revoked, how agent actions are recorded, how incidents are escalated, and whether costs and error rates improve in production.
What to watch next
The next meaningful signals will be a product name or launch timeline, the platforms and agent frameworks supported, and examples of customers using the layer in live workflows. Technical documentation and measurable results would help distinguish a practical enterprise control system from a broad architectural ambition.
TCS’s AI business is growing quickly, but the wider services market remains selective. A governance layer could help enterprises deploy more agents with greater confidence; its value will ultimately depend on whether it can make those systems easier to control, safer to operate and cheaper to run at scale.
