Industrial AI Is moving from automation to autonomy

The Hague (Netherlands): Visualise this: Five alarms go off at once in the control room of a sprawling industrial plant and the young operator scrambles to trace the fault, risking costly downtime. Then, industrial AI steps in by retrieving past responses to similar situations and also suggesting the best course of action in real time. “That’s the new dynamic,” said Cesar Bravo, Director (AI Solutions), Honeywell Process Solutions, in an interview with The Times of India on the sidelines of Honeywell Users Group Europe conference on how AI is reshaping plant operations. Excerpts:

How do you see AI reshaping industrial operations?Bravo: AI has long been part of industrial systems, but the rise of generative AI created an inflection point. Earlier, you needed a data scientist to interpret models; now anyone — an operator or an engineer — can interact with the system using natural language.

That democratization is transformative. We’re seeing real, scalable implementations that create measurable value across industries.How does the AI system actually work in these industrial settings?Bravo: We primarily use machine-learning models built in close collaboration with subject-matter experts at the industrial plant. The process begins by understanding the physical phenomenon: what’s happening, which variables influence outcomes and by collecting years of operational data.

After training, we validate predictions directly with users to ensure they make sense in a real-world context. That human validation is critical to success.What’s the level of AI adoption in India?Bravo: India is one of our major markets and development centers. A large part of our AI engineering and solution design happens there. We’re running multiple pilots with Indian customers with regard to control operations, asset-performance management, and other areas.

There’s strong interest and a willingness to experiment, which makes India a very dynamic space for industrial AI.Which category of industries are leading AI adoption? How is it beneficialBravo: Process industries like refining, petrochemicals, and metals are at the forefront. But we’re also seeing growing applications in life sciences, semiconductors, and even smart buildings like airports where multiple systems like access control, fire safety, and energy management generate massive amounts of data.

The same AI principles apply: analyze, contextualize, and act in real time. A key use case is operator decision support.

When an alarm occurs, our system can review historical data and suggest what previous operators did in similar situations. This preserves institutional knowledge and bridges the skill gap as experienced operators retire.How do you address workforce fears around AI adoption?Bravo: Involving end-users from the beginning is essential. We emphasize that AI is meant to augment, not replace, human capability.

When operators understand that the technology simplifies decision-making and enhances safety, adoption becomes much easier.Cybersecurity is a growing concern in AI systems. How is Honeywell approaching it?Bravo: We design our architectures so that operational data stays within the customer’s premises. Control operations are executed locally, not in the cloud. In addition, we deploy advanced OT cybersecurity systems that continuously monitor for anomalies and threats, using AI to recognize suspicious patterns.

This is far beyond traditional antivirus—it’s proactive threat detection based on years of industrial experience.And what about affordability of AI solutions - especially for small and medium industries?Bravo: Cost is always a consideration, but we design solutions around the value they create. The goal is to optimize processing capacity and architecture so the investment is proportional to the operational gains. Customers today are very pragmatic—they look for a clear ROI, and that’s how we structure every implementation.Finally, what does the future of industrial AI look like?Bravo: AI will become a daily companion for every professional, from plant operators to engineers. Intelligent agents will take over repetitive tasks and provide recommendations in real time. We’re moving from automation to autonomy, where AI doesn’t just assist—it anticipates. That journey has just begun, and the next few years will see it accelerate dramatically.(This correspondent is The Hague at the invitation of Honeywell Process Solutions)