What if factories could think, predict and act before a problem occurs? That possibility is moving closer to reality as physical AI, robotics and connected assets reshape manufacturing. The modernization is no longer about upgrading technology, it is about fundamentally rethinking how factories operate.
In a conversation with Thiruamuthan, Assistant Editor at Industry Outlook, Vijay Srinivasan, Chief Growth Officer, Hitachi Digital Services, discusses the growing role of physical AI in manufacturing, the rise of Lighthouse factories, IT/OT convergence and what it takes to turn enterprise modernization into measurable business value.
With over 27 years of experience across IBM, Cognizant, LTIMindtree, Randstad Digital and Hitachi Digital Services, Vijay has led complex transformation, growth and large-scale technology initiatives across industries and global markets. His expertise spans AI and digital transformation, industrial digital, enterprise growth, strategic partnerships, GCC transformation and large transformation programs.
Read the full interview to understand what it will take for manufacturers to turn AI into real business value.
Q - You've spent over two decades helping enterprises navigate large-scale transformation. What has been the biggest shift in how manufacturers approach enterprise modernization today?
A - One of the key strategic shifts has been the evolution from traditional IT-driven digitization to true physical AI at scale across the manufacturing ecosystem. While many speak about AI, there are still very few companies in the world that are truly operating and engineering solutions in a physical AI environment - which is where our deep IT/OT heritage across 100+ years of industrial manufacturing at Hitachi DS gives us a unique edge.
The biggest shift I've seen is that manufacturers are moving from digital transformation to what I would call physical AI transformation.
For years, enterprise modernization was focused on digitizing processes, upgrading ERP systems, automating workflows, and improving operational efficiency. While those initiatives remain important, today's leading manufacturers are asking a fundamentally different question: How do we create intelligent operations that can sense, predict, decide, and act in real time?
What makes this possible is the convergence of operational technology, engineering systems, enterprise platforms, industrial data, and AI.
In the past, manufacturers primarily used data to understand what happened. Today, they're looking to use AI to determine what will happen, what should happen, and increasingly, to automate the response itself.
This is where Physical AI becomes exciting. Unlike traditional enterprise AI that operates on documents and transactions, Physical AI operates on machines, production lines, assets, energy systems, supply chains, and even entire industrial ecosystems. It combines real-world operational data with AI models to improve throughput, reduce downtime, optimize energy consumption, enhance quality, and increase workforce productivity.
What gives organizations like Hitachi DS a unique perspective is our heritage across both the digital and physical worlds. We understand enterprise technology, but we also understand rail, energy, manufacturing, industrial automation, and critical infrastructure. That combination allows us to connect data from the shop floor through to the boardroom and transform it into actionable intelligence.
The manufacturers creating the most value today are no longer modernizing technology for technology's sake. They're building intelligent, autonomous, and resilient enterprises where AI is embedded directly into operational decision-making.
If you look at Lighthouse factories, for example, Hitachi DS has built World Economic Forum-recognized Lighthouse factories - including flagship deployments in the US and Japan - that are predominantly driven by AI, private 5G connectivity, and advanced robotics, such as Boston Dynamics. Think about what that means for the shop floor. Operations can potentially run 24/7, with autonomous robots taking on mission-critical tasks, including operating in complex or difficult-to-access environments.
So, if I were to summarize the shift in one sentence, it's this: we've moved from connecting systems to creating intelligent systems that can continuously optimize the physical world.
The real transformation is not about digitizing factories. It is about making the physical world intelligent enough to continuously improve itself.
Q - Manufacturers are investing heavily in AI, but not all are seeing the same results. What separates organizations that turn AI into a true growth driver from those that don’t?
A – Many organizations have invested heavily over the years in PLM, MES, SCADA, ERP, historians, quality systems, and supply chain platforms. The challenge is that these systems often operate in silos. PLM knows how a product was designed. MES knows how it was built. SCADA knows what's happening on the production line in real time. ERP knows the financial and supply chain impacts. Yet very few organizations have historically connected these environments into a unified intelligence layer. The companies realizing the most value typically get three things right based on our experiences.
First, they have a strong industrial data foundation that connects IT and OT environments. AI is only as effective as the operational context behind it.
Second, they focus on business outcomes rather than use cases. Instead of asking, 'Where can we use AI?', they ask, 'How can we improve throughput, quality, yield, energy efficiency, customer service, or asset reliability?'
Third, they embed AI into workflows so insights drive action. An AI model that predicts a failure is useful. An AI-enabled operation that automatically schedules maintenance, optimizes production, and prevents downtime creates measurable business value.
At Hitachi DS, we see this every day because we operate at the intersection of digital and physical systems. Whether it's manufacturing, energy, transportation, or industrial infrastructure, the greatest opportunities emerge when AI can understand the operational context and continuously optimize real-world outcomes.
Q - As factories become more connected, how important is IT/OT convergence in turning operational data into smarter business decisions?
A – I think IT/OT convergence is fairly advanced today - in fact, we look at it as a unified ET-IT-OT triad connecting Engineering Technology, Information Technology, and Operational Technology. That is one of the foundational reasons we refer to this as Industry 5.0.
Historically, operational technology systems such as SCADA, MES, PLCs, and industrial control systems were designed to run factories safely and efficiently, while enterprise IT systems such as ERP, PLM, CRM, and supply chain platforms were designed to manage business processes. Both generated enormous amounts of data, but they largely operated in separate worlds.
To make smarter business decisions, organizations need to understand not only what is happening in the business but also why it is happening on the factory floor. That requires connecting engineering data from PLM, operational data from MES and SCADA, asset data from maintenance systems, and enterprise data from ERP and supply chain platforms into a common intelligence layer.
We say that IT tells you what your business wants to achieve, while OT tells you what's physically possible.
What's even more exciting is that IT/OT convergence is becoming the foundation for Physical AI. AI models need context to make meaningful recommendations. By combining signals from the factory floor with enterprise and engineering data, AI can predict failures, optimize production schedules, improve quality, reduce energy usage, and increasingly automate operational decisions.
Also Read: Moving Beyond Cost: How GCCs Are Driving Core Engineering Innovation
Q - Many organizations begin their modernization journey with strong momentum but struggle to scale. Where do most transformation programs fall short?
A - In my experience, transformation programs rarely fail because of technology. They fail because organizations underestimate the challenge of scaling operational change across the enterprise.
Most organizations think change management is about communications, training, and stakeholder engagement after the technology has been deployed. In reality, successful transformation starts with changing how people make decisions, collaborate, and measure success.
Take manufacturing as an example. A company may successfully implement a new ERP platform, modernize MES and SCADA environments, or deploy AI solutions. Technically, the project may be successful, but if planners, production teams, maintenance engineers, quality teams, and plant managers continue operating the same way they did before, the business impact is often limited.
What I've learned over the years is that transformation succeeds when people trust the new ways of working.
That requires three things:
First, leadership alignment. Everyone must be aligned on the business outcomes being pursued, whether that's higher throughput, improved quality, faster product launches, or greater resilience.
Second, operational ownership. Transformation is not an IT project. The business must own the outcomes. The most successful programs we’ve seen have manufacturing, engineering, operations, and IT working as one team.
Third, workforce empowerment. As AI, automation, and Physical AI become more prevalent, the objective is not replacing human expertise. It's augmenting it. The organizations that scale successfully invest heavily in helping employees understand how AI and data can improve decision-making rather than threaten existing roles.
In many cases, the challenge isn't the technology. It's creating confidence that the new operating model is better than the old one. That's why we say that digital transformation is only 30% technology and 70% organizational adoption.
What’s your personal mantra?
"Stay curious, stay humble, and stay focused on outcomes."
.jpg "Industrial Packers: Engineering Industrial Packaging for India's Manufacturing Future")
We use cookies to ensure you get the best experience on our website. Read more...
