AI can help factories find bottlenecks before production stops: Tata Communications’ Praveen Arora

When a factory runs below capacity, the first line of questioning tries to answer ‘where production slows down’. Today, by connecting data that different machines and systems already collect, AI can help answer it, believes Tata Communications’ Praveen Arora.

Speaking to ET AI at the India Mobile Congress (IMC) 2026, Arora, vice president of the firm’s Internet of Things business unit, said manufacturers can use that information to identify bottlenecks, improve output and reduce operating costs.

Finding the gaps in factory output


Factories have collected large data through sensors, equipment, applications and production processes, but their systems have often worked separately, Arora said, explaining that combining those inputs gives operators a clearer picture of how their plants perform.

“You get a complete visibility of where is the bottleneck. You know what is your installed capacity versus the gap,” he said.

Arora said Tata Communications approaches these deployments with goals such as improving efficiency, saving energy or making workers safer instead of treating each piece of technology separately.

“The way we have evolved is we work with the business outcome in mind,” he said, describing how the company combines connectivity, connected devices and AI tools to support those goals.

The company provides networks, devices and sensors that feed information into its applications and platform, Arora said, adding that it chooses between three aspects: processing data near the equipment, using the cloud according to the application and the overall cost of running the system.

He said its Factory 360 offering brings together data from sensors, applications and factory control systems, allowing operators to spot changes in performance before an equipment breaks or a production line stops.

Conventional alarms have largely required operators to react after a problem appears, while AI can detect early signs of something going wrong and feed that information back into operations, Arora said.

“Before you realise that something is going wrong, you are detecting advanced signals of anomaly and taking it as a feedback loop,” he said.

Arora linked these operational gains to worker safety, saying that preventing incidents can also help keep production lines running, while information from different systems helps operators act on problems earlier.

“So if I'm taking care of my workers and safety, I'm taking care of efficiency because the lines are not stopping. If I'm going for efficiency, one of the factors in the levers is definitely the safety as well, out of the 10 other levers,” he said.

Starting with existing machinery

For smaller manufacturers with limited budgets, Arora said cost and efficiency should be the starting point, with many plants already collecting sensor data that can be brought together and analysed.

Even machines that are 20 or 30 years old can contribute to such systems, he said, explaining that converters can turn their analogue signals into digital information and connect them to a broader view of manufacturing operations.

Arora said the company combines inputs from existing devices, checks for security gaps and applies analytics to identify efficiency improvements, while manufacturers can test the approach through a pilot using data from their existing systems.

The algorithms can also run in a customer's premises without sending data outside, he said.

“So you can do a proof of concept and that can be shown in less than a week,” Arora said, referring to demonstrating the approach through a trial.

He said the benefits also extend to workers who previously spent time identifying the cause of a problem and deciding how to resolve it, with AI providing data in a more intuitive form to help them act faster.

Arora cited throughput improvements ranging from 30% to 200% in some situations while using the same resources.

He described the role of AI as equipping employees to make decisions, “It's about equipping the worker or the employee with more information to make decisions.”