General Electric has played the part of industrial monolith for so many of its 126 years that its Intelligent Platforms division is enough to make at least one corner of the company feel like an honest startup, especially in regard to its efforts in automated intelligence. The division has acquired more than a dozen smaller companies during recent years, an "inorganic acquisition spree," to use the words of chief of strategy Rich Carpenter. Among the highlights: Intellution, an industrial software company, and Mountain Systems, which allowed for an increased reach in manufacturing.
"We see innovation almost running rampant right now, and you can't always tell what's going to succeed and what's not," Carpenter said. "We're looking across the board at promising technology, backing some startup ventures and making use of some of those products in our own initiatives."
Automated intelligence is among those initiatives. It's a focus for Carpenter, an Industrial Internet champion who wants to make factories smarter. Why not provide machines with the same ability we have to turn to the Internet to make better decisions, and to draw on a broader set of variables in order to make more profitable decisions?
"I think we're trying to change the nature of the work that people have to do on the plant floor," Carpenter said. "Today, a lot of people have their hands full just keeping the plant running."
We're looking across the board at promising technology, backing some
startup ventures and making use of some of those products in our own initiatives.
— Rich Carpenter
Free those hands and more people will have more time to transition upward with advanced tasks, like analytics. "The primary problem," Carpenter said, "is that 90% of the data that people have never gets used." That's why a new breed of Six Sigma engineers is working with data that Carpenter describes as being like oil -- it has value, and that value has to be extracted -- and then mined for correlations. When the engineers discover those correlations, "We can put analytics in place and advise people long before problems actually occur."
Carpenter paints his ideas in incredibly big pictures with incredibly small details. He wants to model equipment on good behavior. He wants to model people on good behavior, too. Who are the best people with the best results? How are they accomplishing those results? How can their processes be captured? And how can others be advised and trained to do the same?
These are questions that can be answered when more than 10% of data is actually crunched, he said. "Moving from that reactive 'What happened?' to seeing something coming, it's a big change in skill set, a big change in approach and a big change in environment."
