How Tata Steel Uses AI To Deliver Steel In Hours

Tata Steel has leveraged AI to analyze workforce safety data – identifying critical safety insights including root causes of recurring incidents, high-risk zones, and critical safety gaps. These insights helped identify critical high risk high impact zones across plants. For such zones, the steel giant has incorporated a Gen AI based platform that analyses real-time video feeds to perform a 360-degree situational analysis. This solution considers the complete physical and operational context to identify hazards and notifies the line manager immediately. Additionally, the system auto-learns with human feedback to continuously enhance its operational understanding. 

For better customer experience, Tata Steel has deployed a multimodal, multilingual Agent to act as a dual-layer problem solver. First, it utilises internal knowledge bases to autonomously solve routine queries without human involvement. Second, for actual complaints, it identifies the specific problem by analysing user provided inputs - including text descriptions and image artefacts to detect defect types and auto-assigns the issue to the correct resolver group within Tata Steel. This automated triage solves the previously existing routing bottleneck, ensuring the complaint reaches the exact team required for immediate action resulting in faster resolution times and an enhanced customer experience. FE CIO speaks with Jayanta Banerjee, Group CIO, Tata Steel, about the company’s digital transformation priorities.

Edited Excerpts:

Tata Steel is extensively using AI. What’s your take on measuring Return On Investment (RoI) from AI. What benefits have you achieved from it?  

The digital transformation initiatives are tied to the business key performance indicators (KPIs). If the business KPI has not moved, then there is no value in technology. If the KPI has moved, it is irrelevant whether I use Gen AI or AI because it is always a combination. There is nothing that works in isolation, all systems have to work together. There is a process and then technology, the infrastructure, cloud, involving the data repository. Then there is a mathematical and Generative algorithm on top of it.

So how do you look at the benefits from AI from a qualitative sense ?

Scalability is one of the most important benefits of AI, beyond just quantitative or qualitative gains. The 600 models we run in India today, including both narrow and generative AI, help us solve problems at multiple levels, across functions and hierarchies, and the scalability and democratisation that Generative AI enables is significant.

Equally important is how AI is qualitatively reshaping work culture and work ethic, which is ultimately more transformative, with the quantitative benefits realised along the way.

In global forums such as the MIT Industry Academy Board for AI and a World Economic Forum platform, a key question we discuss is: What happens when Generative AI can do much of what humans do? The emerging view is that lower‑order, more transactional roles will increasingly be taken over by AI, and humans will move to higher‑order work.

As GenAI takes on routine tasks, people step up to higher‑faculty roles; over time, AI learns from that and also moves up, prompting humans to climb further. The real impact of AI, therefore, is not just cost efficiency but freeing humans from transactional work so they can focus on more complex, creative and strategic problems.

Which areas are currently seeing improvements at Tata Steel that can be directly traced to Artificial Intelligence ?

Workplace Safety is the area where Tata Steel has received the maximum results from AI. It is  video image-based. Safety is not measurable because the results cannot be expressed in terms of whether there were less or no accidents or fatalities at the manufacturing sites. However, AI is helping in generally pre-empting a lot of accidents. 

Logistics is another area where, through sensors, Generative AI andML, we are able to see which truck is not moving in the fleet and, hence, my shipment will be delayed. I can proactively pre-empt a customer complaint. 

Generative AI is applicable in insights generation, especially where there is a lot of unstructured as well as structured data. There is a lot of data in financial analytics, competitive analytics, scenario building, strategy, and planning—which is a mix of unstructured and structured data. The structured data is in SAP. Similarly, HR data is largely unstructured. 

Using AI, my meetings have reduced drastically. The Tata Steel Digital Assistant (TDA) helps me to prepare for the meetings. I have my questions rehearsed well in advance. Without the need to go through a 100-page document to come up with a question. The AI is doing it for me and giving me insights. There is always a human judgment on top of it.

How has the final customer benefited from the AI interventions at Tata Steel ?

In the context of the steel industry, we have an aspiration to become like a Zepto or a Blinkit.This does not mean we will deliver steel in 10 minutes. But we can deliver steel in the range of about 72 hours on the Aashiyana Platform.

Traditionally, steel customers would place a bulk order and then wait with little visibility. At Tata Steel, the process is app‑driven: once an order is placed, AI‑enabled systems check specifications, grade availability, inventory and line capacity to optimise the supply chain and provide a clear confirmation with committed lead time.

Like other e‑commerce platforms, customers can track their consignment status, fundamentally changing the steel buying journey. On Aashiyana, where average ticket sizes are high, customers often drop off early, so we use an agentic AI framework to send intelligent nudges, gradually moving them forward, generating “dynamic data” and feeding insights back into the experience.

By creating data on the move and using it to personalise and de‑friction the journey, Tata Steel has been able to steadily grow revenue from the Aashiyana online portal quarter after quarter.