Manufacturers are constantly juggling the need to automate with an understanding for the best ways to implement these new and exciting technologies. The worst case scenario being huge investments that alienate the workforce and fail to deliver the efficiency and production enhancements that were prophesied.
The proper mix is specific to each manufacturing environment, but I recently sat down with Rahul Negi, an Industrial AI Leader at Honeywell, to discuss strategies for balancing the financial, technological and practical elements of integrating additional automation assets.
Jeff Reinke, editorial director: The momentum behind automation implementations in manufacturing has not slowed. What do you feel are some of the biggest missteps that companies are taking in their eagerness to put these new technologies to work?
Rahul Negi, Honeywell Process Automation: One of the biggest missteps is that companies often overlook just how time and resource-intensive technology deployment can be. The proof is evident in today’s global supply chains, energy systems and critical infrastructure – all of which face unprecedented strain.
For example, new research reveals AI intensifies work before efficiencies materialize, and studies continue to affirm AI’s costly energy footprint. As a result, companies face shortages, outages and delays that are tough to recover from.
Another misstep we’ve observed in automation implementation is that companies don’t take the time to understand how different data points interconnect within and impact their operations. Data is often disconnected and as a result, offers limited adaptability and slows optimization. Building on this, global industrial labor shortages hinder data interpretation and action on outcomes. All of these challenges compound in catastrophic cybersecurity, IT and OT risks that can take months, if not years, to reverse.
The good news is that we’ve seen what drives progress. Working with an array of customers and partners has shown us that physical AI improves business outcomes by enhancing safety, connecting data and improving decision making, all while supporting workers who may not have decades of experience.
JR: A common perception is that automation replaces people, and therefore jobs. How would you recommend introducing these new initiatives to the workforce while quelling concerns over being replaced?
RN: Companies must keep their focus on improving human productivity rather than replacing roles. We’ve seen how quickly technology can help human workers improve their roles and take on new responsibilities or positions, but as a result, they fear that technology will eventually wipe out their job.
While these concerns are plausible, they don’t reflect how industrial automation operates. Industrial automation always requires human oversight – humans are the catalyst for better outcomes due to their specialized knowledge, contextual understanding and judgment that ensures safe and beneficial deployment.
At Honeywell, we believe companies have a collective responsibility to equip today’s workers and leaders with the tools and knowledge they need to navigate the path to autonomy and empower the workforce to feel confident in how they deploy automation. Internally, we’re infusing AI into Honeywell systems like Experion Process Knowledge System (PKS) – an alarm assistant for operators that reviews and analyzes all historical data to provide recommendations for operators to address risks and issues in real time.
JR: If you had to pinpoint one driving force that should fuel all automation strategies and implementations, what would it be?
RN: If you asked me this question a year ago, I likely would’ve said it’s the convergence of three technologies: AI, 5G and the cloud. But now these technologies are enabling a new, emerging lever – physical AI.
Building on my earlier point about physical AI being the cornerstone of successful automation implementation, I’ve seen firsthand that physical AI leads to greater efficiency, preventative maintenance that reduces downtime and helps close the skills gap. It does so by providing workers, who have limited experience, with the ability to function as if they’ve been on the job for years – all of which leads to real productivity gains.
JR: There’s an abundance of tools and providers in the marketplace right now. What selection advice would you offer in sorting through all these options?
RN: The market is flooded with AI tools and providers, but what we’ve learned is that deep domain expertise matters more than anything else. For those sorting through the vast sea of options, I recommend seeking those who understand how automation can be applied to OT across sectors like energy, manufacturing, utilities and industrial and commercial buildings.
JR: What is your take on the ideal roles and overall impact of AI? Are AI’s ideal applications really understood?
RN: We’re starting to see more clarity around where AI delivers value, because we’ve entered a stage where leading use cases have moved from experimentation to scaled deployment. Industrial facilities like oil refineries and petrochemical processing plants are generating vast operational datasets daily, from thousands of endpoints or assets.
These environments are where AI-powered autonomy thrives. For example, we have collaborated with ADNOC Borouge to develop AI-powered autonomous operations. According to ADNOC Borouge, these solutions once implemented across their sites, have a potential to improve efficiency gains by 20 percent, reduce downtime by 20 percent and reduce operating costs by 15 percent. This is a prime example of AI’s idea role – to extend decision-making beyond individual plants to the entire enterprise.
But historically speaking, reaching full autonomy has been challenging because data was siloed and traditional compute power couldn’t handle the influx of data that facilities were producing. Agentic and physical AI are changing that – they’re processing these huge datasets in the cloud and on edge.
JR: Outside of AI, what do you see as some of the biggest trends impacting manufacturing overall in 2026?
RN: There are three key trends that will have material impact on the manufacturing industry this year.
First, the continued exodus of experienced operators – a challenge that we’ve seen worsen in recent months, with 26 percent of the national industrial workforce eligible for retirement – who have decades of experience and are retiring, taking their knowledge with them.
New operators face a steep learning curve and struggle to manage the peak performance of their plants and assets. This has significant implications for manufacturing enterprises, because new headwinds result like unplanned downtime, rising maintenance costs and aging infrastructure. To address these issues, manufacturers are leaning on AI to support new operators in sharpening their skills and maximizing their productivity.
Another trend we’re witnessing is the rise of API-first approaches and interoperability. Our customers are increasingly asking for consistent user experiences and interfaces.
And lastly, we’ll begin to see manufacturers achieve full autonomy with more open solutions and ecosystems, embedded knowledge and domain expertise, and value-first deployments. There have been many challenges in making the shift to autonomy over the past year, like customers getting stuck in pilot purgatory, but they’re moving past this stage and seeing time to value accelerate.
Solutions are now being delivered in a timelier manner and end-to-end lifecycle management is propelling the path to autonomy.
Rahul Negi is an executive at Honeywell Process Automation leading Digitalization, Autonomous, and AI initiatives, with over 20 years of expertise in Strategy, Consulting, Business Development, and AI/ML.
