A new generation of software-defined automation is quietly changing the character of modern manufacturing. By decoupling control from hardware, it is turning rigid systems into fluid networks of intelligence, making operations more adaptable, efficient, and ready for the age of AI.
The history of industrial automation is a series of redefinitions of control. The 1970s marked the first great leap forward, when pneumatic systems were replaced by digital controls. The early 2000s brought a second transformation with the arrival of Windows-based architectures and Ethernet networks. And now, a third wave is building, one that is more philosophical than mechanical. It is about loosening the rigid bond between hardware and software, letting intelligence move freely across the plant floor and, in the process, changing how manufacturers evolve.
Joe Bastone, Director and Growth Initiative Leader at Honeywell Industrial Automation, has lived through two of these eras. His career began in the era of CRTs and hardwired controllers and has carried him into an age when industrial control is starting to look more like software engineering than mechanical maintenance. “The earlier revolutions swapped one technology for another,” he explains. “This one unbinds them. We are separating the software logic from the hardware so we can adapt faster, scale smarter, and build systems that never stop improving.”
Moving away from the monolith
For decades, automation systems have been built like cathedrals, precise, intricate, and almost impossible to alter without enormous cost and risk. Once commissioned, they were expected to run for decades with minimal change. Every update was a project, not an event. That thinking is finally beginning to fade.
Software-defined automation replaces the idea of a system as a single structure with something far more fluid. The software can now live independently of the equipment that runs it. Control functions, once tied to specific pieces of hardware, are now packaged as portable modules, containers that can be redeployed, updated, or extended without dismantling the entire system. “It is like moving from rebuilding a bridge every time you need a new lane to simply adding the lane,” Bastone says. “Once you separate control logic from the metal, innovation stops being disruptive.”
The change sounds subtle, but its implications are enormous. Manufacturers can now add capabilities in weeks instead of years. New features or optimisations can be introduced without shutting down production. In some plants, this has already transformed the relationship between engineering and operations, enabling continuous improvement in real time rather than through carefully scheduled upgrades.
The slow collapse of the data wall
Even with flexible systems, manufacturers still face an older problem: data isolation. The information produced by their processes is often spread across incompatible systems and formats. It is a patchwork of sensors, PLCs, and control applications that were never designed to speak to one another.
“The typical site has dozens of data silos,” Bastone says. “In one refinery, we counted over a hundred and sixty separate systems, all collecting data, none of them sharing it. It is not that the information does not exist; it just has no way to connect.” This fragmentation limits visibility and makes it hard to apply advanced analytics or artificial intelligence at scale. Software-defined automation begins to solve that by building a connective layer, a kind of industrial data fabric that links existing systems and contextualises their data without forcing manufacturers to start again from scratch.
The approach changes how insight is generated. Maintenance teams can match sensor data with production schedules to predict downtime. Supply chains can align shipments with live output data. Energy consumption can be tracked against performance and yield. Once data becomes connected and contextual, it turns into something much more powerful: foresight.
Escaping the endless pilot
Bastone describes another recurring frustration across the industry in pilot purgatory. It is the place where innovation goes to die. New technologies are tested, proven, and celebrated, but never scaled. Each pilot sits in isolation because the underlying systems are too different, too rigid, or too expensive to replicate.
“The issue was never the lack of good ideas,” he says. “It was the lack of a framework to repeat success. You cannot industrialise digital transformation if every site starts from zero.”
Software-defined automation changes that equation. Once an application is built within a standardised, modular environment, it can be duplicated anywhere. The value is not in the individual pilot but in the ability to repeat and adapt it. A predictive maintenance algorithm developed for one plant can be rolled out to ten others with minimal integration work.
For manufacturers used to years of stalled projects, that repeatability is quietly revolutionary. It transforms innovation from a proof of concept into a habit.
The edge, the cloud, and everything between
One of the most common misconceptions about software-defined automation is that it pushes everything into the cloud. In reality, it gives manufacturers the freedom to place. Processes that require sub-millisecond latency stay on-site, while analytics and training models can live in the cloud, where computational power is abundant.
“It is not a question of edge or cloud,” Bastone explains. “It is about having a continuum that connects them. Some workloads make sense to run alongside the process; others make sense to run centrally. What matters is that they share the same structure so they can communicate seamlessly.”
This flexibility means that industrial data no longer needs to choose between responsiveness and scale. Edge systems handle control, safety, and immediate response; cloud systems handle optimisation, modelling, and AI. Together, they create a feedback loop in which every decision feeds into the next.
As that loop strengthens, the plant begins to behave more like an organism than a machine, aware of its own performance, capable of adjusting to change, and increasingly self-improving.
People, process, and cultural inertia
The shift to software-defined automation is not purely technical. It requires a rethinking of skills, roles, and responsibilities. Control engineers now work alongside software developers, data scientists, and IT specialists. That convergence can be unsettling, but it is also necessary.
“When I started, we were worried about wiring diagrams,” Bastone says. “Now we are talking about Kubernetes clusters and container orchestration. The terminology has changed, but the goal is the same: keeping the process stable and safe while giving people the tools to innovate.”
This cultural adjustment is as challenging as any technological upgrade. Operational teams, trained to value stability above all, must now accept that constant iteration does not mean instability. It means adaptability. IT teams, on the other hand, must learn that uptime in manufacturing is measured not in hours but in lives of equipment and processes.
The success of the transition depends on building trust between these groups. It demands leadership that understands both sides, technical precision and organisational psychology.
Building for evolution, not replacement
Honeywell’s Aptica platform is a glimpse of how software-defined automation works in practice. It allows manufacturers to connect existing systems, deploy new applications as modular components, and gradually expand capabilities. The first step might be as simple as digitising paper batch records. From there, companies can layer on analytics, then introduce predictive models, and eventually move towards semi-autonomous operation.
This incremental approach reduces risk. Instead of betting on a full-scale transformation, manufacturers can evolve piece by piece, proving value at every stage. The result is not disruption but renewal, a steady accumulation of capability that makes each system more innovative and more efficient than the one before. “You are not throwing away your history,” Bastone explains. “You are building on it. The plant becomes a living system that grows with every improvement.”
The financial benefits of this modularity are already being felt. Plants that have adopted the model report lower maintenance costs, faster updates, and measurable energy savings. Yet Bastone insists that the actual value is strategic. Once architecture becomes software-defined, it never needs to be replaced again. It just evolves.
The philosophy of continuous change
Beneath the technology lies a more profound shift in how manufacturing thinks about time. In the old world, automation systems were designed for endurance, installed once, upgraded rarely, and retired after decades. In the new world, they are designed for movement. Change is no longer an interruption; it is the operating model itself.
This mindset aligns naturally with artificial intelligence. AI thrives on iteration, on learning from data and refining itself. Software-defined automation provides the infrastructure to support that learning. It ensures that AI can access real-time information, influence control strategies, and scale across plants without rebuilding the entire system each time.
It also changes the conversation about sustainability. Smarter systems consume less energy and can adjust output to match demand, reducing waste. They can predict failures before they happen and extend asset lifespans. Efficiency and environmental performance become two sides of the same architecture.
Towards intelligent industry
The transformation that Bastone describes is gradual but inevitable. It represents the industrial world’s migration from fixed-function control to flexible intelligence. Factories will not look different at first, but their behaviour will. They will be able to adapt on the fly, scale innovation instantly, and integrate AI directly into their operations. “The most important thing about this change is that it removes the waiting,” Bastone says. “You no longer must plan the next revolution. You just evolve continuously.”
For an industry built on precision, reliability, and caution, that statement might sound radical. Yet it reflects a growing truth: that the systems keeping the world’s plants running can no longer afford to stand still.
Software-defined automation is not an abstract concept—it is the quiet architecture of the future. It represents an industrial world that can learn as quickly as it produces, where every process is connected, and where intelligence is not something added to the system but something that defines it. Manufacturing has always been about control. What is changing now is who, or what, exercises it.
