🚀 The Missing Link Between AI Strategy and Operational Reality

[Hasan S. Cemkan]

🤔 AI Investments in Manufacturing: Where Are We Heading?​

Manufacturing leaders are facing increasing pressure to invest in artificial intelligence and ensure these investments deliver measurable business value. According to a PwC survey, 81% of over 500 executives plan to increase their AI spending in the next three years. This raises a critical question for senior leaders: Is the next AI project advancing the organization toward a scalable strategy, or is it merely responding to the pressure to "do something" with AI?

🚧 Without an Operational Foundation, AI Remains a Dream​

For many manufacturers, especially those operating in older facilities, the real challenge isn't finding a potential AI application. It's building the operational foundation required for that application to produce reliable results and eventually scale.

Before investing, leadership needs to define the desired business outcome. What problem are we trying to solve? Is AI the right technology for this problem? And if so, how does this investment fit into the facility's long-term strategy?

These questions quickly lead to a topic that has historically received much less attention in the boardroom: operational technology (OT).

🔗 AI Strategy = OT Strategy​

An effective manufacturing AI strategy is inseparable from an OT strategy. AI readiness begins with connectivity. Equipment must be networked for the necessary operational information to be available.

But simply connecting machines isn't enough. Data from manufacturing systems also needs to be cataloged, contextualized, and normalized so that actions and process states carry consistent meaning.

For example, consider two machines performing related functions on the same production line. Experienced operators might understand that different labels, naming conventions, or signals essentially represent the same event. An AI application does not inherently possess this institutional knowledge.

If one machine describes an action one way, and the next machine describes the same action differently, the system needs to be given enough context to understand that relationship. Therefore, collecting factory floor data is not the same as making that data usable for AI.

This becomes particularly challenging in older facilities, where plants may contain decades of equipment from different manufacturers, multiple generations of control systems, and varying levels of automation. Very few facilities were designed from the outset around a single data model or network architecture.

This doesn't mean manufacturers should wait for every asset to be modernized before leveraging AI. It means leadership needs to have a clear picture of what's ready today and what needs to change for AI to advance further in operations.

🎯 Start with the Right Problem, Not Necessarily the Biggest Problem​

Valuable applications for AI in manufacturing exist today. The key is to match the ambition of the application with the readiness of the environment.

Some of the most powerful early opportunities can be intentionally narrow: analyzing high volumes of operational information to identify patterns, supporting business intelligence, recommending improvements, or applying AI within a well-understood island of automation.

Expecting AI to control machines or manage processes across an entire manufacturing environment is a much larger leap.

A good example of the more targeted approach comes from a process chemistry application E Tech Group developed to help scientists optimize experiments using machine learning.

SCADA-integrated dashboards built in Ignition connected lab automation equipment, analytical instruments, and cloud-based machine learning models in a centralized environment.

Scientists could manage experimental campaigns, monitor model-suggested experiments, and capture results with far less manual tracking. The solution improved traceability and reproducibility and established a foundation for future closed-loop autonomous experiments.

The point isn't that every manufacturer should pursue this specific application. The point is that AI was applied to a clearly defined problem in an environment where the necessary connectivity, data, and process context could be established.

Manufacturers can adopt the same approach on the factory floor: Identify where infrastructure and data already support a valuable use case, drive success there, and continue building the foundation required for broader applications.

🚫 IT Applications Cannot Be Directly Transferred to OT​

Building an AI-ready foundation requires another important distinction: IT and OT environments are not interchangeable. Network principles may overlap, but OT networks support physical production processes where availability, predictable communication, and production continuity create very different operating requirements.

For example, an IT-centric monitoring tool might actively scan connected devices and request information across a network. Applying the same methodology to an OT environment without understanding its architecture can inject traffic onto networks already carrying time-sensitive production communications.

E Tech Group has encountered situations where monitoring approaches brought into the plant from the IT environment contributed to intermittent latency and production issues. The technology itself wasn't necessarily the problem; the problem was applying IT methodology without accounting for its downstream impact on OT operations.

As connectivity expands, IT and OT teams need to work together. Principles may be similar, but execution in a manufacturing environment requires OT experience, especially when changes need to be made while production equipment remains operational.

Cybersecurity risks further amplify this. In a food and beverage facility, an earlier IT/OT assessment had identified network segmentation as a priority. When a subsequent enterprise cyber incident occurred, that segmentation helped isolate production, allowing manufacturing operations to continue.

The lesson extends beyond cybersecurity. Jobs that once seemed like backend infrastructure (network segmentation, firewalls, OT architecture, and asset visibility) now directly impact a manufacturer's ability to support broader digital and AI initiatives.

📈 A Successful Pilot Project Doesn't Prove Scalability​

Manufacturing AI adoption is progressing rapidly, but adoption and maturity are not the same thing. Smart Industry also reported that a Rootstock survey of 520 digital transformation leaders found that approximately 94% of manufacturing organizations are using some form of AI, but most applications remain in experimental or pilot phases.

This distinction is important. A successful AI pilot project proves a use case. It does not prove that the facility is ready to scale it. One of the biggest challenges in moving from pilot to production is accounting for edge cases.

While AI can be incredibly powerful, it can still miss contextual conditions that are obvious to an experienced human operator. Humans spend years learning what "normal" looks like in a manufacturing environment. They notice anomalies almost instinctively.