Industrial processes are massively complex and generate massive amounts of data. Anyone who has dug their hands into a manufacturing project can attest to that. Industrial operations folks—I consider them marvels—have traditionally used their genius to run these complex networks, optimizing and improving processes on the basis of their experience. But digital transformation and Industry 4.0 are bringing modern approaches to data analysis. And the enterprises that best marry this old-school intelligence with that of the artificial kind (AI) will end up looking the smartest.
AI can be applied to industrial processes for handling the millions of data points being generated by manufacturing environments and predicting future outcomes of these processes. This we know. This is increasingly at play in myriad industrial settings around the world. For example, in high-volume environments, it is not humanly possible for manufacturers to inspect the quality of each part. They must rely on random inspection. As a result, quality becomes jeopardized, scrap rates increase and the customer experience is affected.
With AI, on the other hand, quality defects can be predicted before the part is produced, saving millions of dollars in unnecessary scrap and warranty costs for the manufacturer.data. Anyone who has dug their hands into a manufacturing project can attest to that. Industrial operations folks—I consider them marvels—have traditionally used their genius to run these complex networks, optimizing and improving processes on the basis of their experience. But digital transformation and Industry 4.0 are bringing modern approaches to data analysis. And the enterprises that best marry this old-school intelligence with that of the artificial kind (AI) will end up looking the smartest.
Consider another case, in which quality metrics for a particular process—welding—can be inconsistent, as the robotic welding station is being fed by different operators. In order to have a consistent weld, AI can be used to drive welding parameters, understanding and predicting when welds go out of spec and ultimately causing less waste, improved quality and asset optimization for the welding robotic itself.
Whether we’re talking about enhanced quality control or improvements with yield or reductions in process time, data analysis can be the differentiator. And that’s where AI can supercharge outcomes.
If you’re wondering where to get started, AI is best applied to areas with repeatedness, where an automation can be instituted and standards can be set. In situations where you’d previously had an operator managing a process, you can leverage AI (which is not reliant on operator changes) to tackle the complex series of steps in order to reach the desired outcome.
Modern AI is used to create predictions that are constantly learning from the process data—a level point of quality with a consistent output not humanly possible otherwise. Yields meet targets. Quality is upheld. In short, AI can transform production processes.
Efficiency and cost is where most businesses start the process. (Optimizing production is often the following step.) Naturally, enterprises don’t always know what their greatest problem is. The devil is in the details, right?
It’s all about asking questions and helping uncover opportunities in which there’s a real need (e.g. problem statement) for optimization, there’s good data, and there’s the right support internally.
We stress the value in developing a problem statement (THIS IS WHAT OUR BUSINESS NEEDS TO FIX), then mapping data to that problem statement. If you don’t start there, you can’t really determine what it is that the data is telling you.
Once the process is begun, of course, data overload can quickly become an issue. Business leaders wonder if they’re collecting the right data. They wonder if the data is in the right place and formatted properly for use. At this point, it’s a simple data-transformation exercise to determine where the data is coming from, how it is being stored, and how insights from the data are being applied to the assets. This is often the most enlightening part of our work, as customers learn the most about themselves. This really drives to the heart of what enterprises are trying to gain from their long-term IoT / Industry 4.0 initiatives.
The first genius is human, I like to say. The second genius is AI. The former excels in strategic decision-making. The latter is the hero of repeatable standards.
A harmonious partnership of intelligences is most productive; I consider the human brain a vital component to artificial intelligence. And vice versa.
