Ask a room full of manufacturers where artificial intelligence will deliver the greatest value, and most will point to enterprise systems—dashboards, planning tools or supply chain platforms. It’s a reasonable guess. It’s also the wrong one.
The work that will determine whether autonomy is achieved actually happens a layer down, in the controllers, sensors and actuators that turn a decision into action. The control layer is the foundation autonomous operations are built on, and it’s where tangible business value is unlocked.
It’s also the layer many manufacturers overlook. Explore four common myths about autonomous operations and how the control layer determines whether autonomy scales across the enterprise.
Myth: AI transforms a plant from the top down
Often, manufacturers invest in AI and expect operations to become autonomous. But that only happens when predictions and optimizations can actually be acted on, which requires a control layer that can sense accurately and execute reliably.
A 2026 Manufacturing Dive article reinforces this point, noting that manufacturers are increasingly turning to sensor technologies to build the foundation for digital transformation.
That’s not a reason to slow down AI investment. It’s a reason to build from the bottom up, starting at the control layer.
Myth: Robots deliver the greatest value
When picturing factory automation, most imagine a robot arm. And there certainly are many robot arms in use—the International Federation of Robotics reports more than 4 million industrial robots in operational use worldwide. However, while that’s a big number, it’s dwarfed by the installed base of industrial controllers, which is almost a billion.
Almost every one of those controllers runs on logic written by hand. Massimiliano Moruzzi, CEO of the physics-based AI firm Xaba—which is backed by Hitachi Ventures—estimates that roughly 80% of automation cost comes from manually developing that controller logic.
The greatest ROI doesn’t come from robots. It comes from leveraging AI to generate and adapt the logic that drives the millions of machines already on the floor.
Myth: AI on the plant floor will be a chatbot
The version of AI most people are familiar with is a chatbot. However, that’s not what belongs on a production line.
Physics-based AI doesn’t read a manual. It learns the physics of a process from real signals—force, vibration, voltage, temperature, and load—and uses them to generate and adapt control logic. For example, a packaging line can tune its own motion profile to cut energy use, or a retrofit can make a 15-year-old machine adaptive without anyone tearing it out and starting over.
A good engineer writes code from experience. A system like this writes from far more process data than any one person gathers in a career, and it keeps learning after commissioning. This allows manufacturers to move from reacting to anticipating.
Myth: Autonomous means unmanned
Autonomy moves operators up the value chain—from running equipment to overseeing systems that increasingly run themselves. As operations mature from reactive to descriptive to predictive to prescriptive, the human role elevates at each step instead of disappearing.
The goal shouldn’t be removing people. It should be to effectively anticipate problems and improve judgment.
The foundation that unlocks ROI
While most of the attention around industrial AI goes to the work above the control layer, this foundation is critical to delivering tangible ROI.
Without the ability to sense and execute, none of the promises of industrial AI can be fulfilled.
Competitive advantage won’t be won in the cloud or on a robot’s spec sheet. It’ll be won in the control layer. Start building the foundation for industrial AI today. Reach out to our team of experts.
