Tata Steel deploys over 300 AI agents in nine months with Google Cloud

Unified data and agentic AI stack moves steel major from experimentation to real-time execution across operations, finance, HR and the shop floor.

Tata Steel, a global steel manufacturer and part of the Tata Group, has deployed over 300 specialised AI agents across its global operations in just nine months as part of an expanded partnership with Google Cloud.

The deployment spans operations, finance, human resources, and shop floor environments, marking a shift from fragmented digital initiatives to a unified agentic AI framework.

The company has built a consolidated data and AI stack on Google Cloud to enable real-time decision-making across its value chain.

This integrates operational data with AI-driven workflows, allowing teams to act on insights as part of daily processes.

AI is no longer a technical experiment but a partner for every employee, said Jayanta Banerjee, CIO, Tata Steel, highlighting how the system allows teams to act on insights instantly.

The use cases span predictive asset maintenance and faster customer response cycles, pointing to a broader shift toward autonomous operations.

Focus on unified data layer and agent-led execution

A key part of the deployment is Tata Steel’s data consolidation strategy, which enables AI systems to operate across structured and unstructured data sources.

The company has built two internal platforms to support this. It includes Zen AI and the Tata Steel Digital Assistant (TDA).

Zen AI functions as a low-code environment that allows developers, engineers, and business teams to build, test and deploy AI agents.

It is built using Google Cloud’s Agent Development Kit and integrates with data platforms such as BigQuery and Cloud Storage to process operational and enterprise data.

The platform brings structured operational data with unstructured inputs such as video and documents within a governed framework.

It enables teams beyond central data science functions to create and deploy AI use cases, expanding adoption across business units.

Complementing this is TDA, a unified decision interface that aggregates data across public sources, enterprise systems and proprietary datasets like call recordings and PDFs.

The system enables employees to query across these layers in real time, combining external signals such as geopolitical sentiment with internal operational data to generate predictive insights.

The system is already being used to handle routine workflows.

Tata Steel said TDA resolves over 70 percent of HR helpdesk queries, reducing manual intervention. In finance and compliance, AI agents are being used for invoice processing, GST classification, and contract analysis.

On the operations side, AI agents help improve safety and equipment performance.

Safety EyeQ, a specialised agent, analyses live video feeds in high-risk zones to detect SOP violations and potential hazards, triggering immediate alerts.

Asset Sphere monitors equipment health to enable predictive maintenance and reduce downtime.

In customer operations, AI agents analyse complaint data, including images, to identify issues and route them to relevant teams. The company said this has reduced turnaround time by up to 50 percent.

The deployment runs on Google Cloud Run, which allows the system to scale based on demand while optimising costs. Tata Steel is also using multiple AI models through Google Cloud’s platform to support different use cases while maintaining governance.

The initiative reflects a shift in enterprise AI adoption, where organisations are embedding AI into core operations and moving from experimentation to continuous, real-time execution.