The beginning of the Industrial Internet of Things (IIoT) has been traced back to 1968, when Richard Morley invented the programmable logic controller (PLC) that General Motors used to make automatic transmissions. Morley’s work unofficially ushered in a new era of industrial connectivity.
In the more than 50 years that followed, nearly every aspect of the manufacturing industry has become connected. Until recently, the flood of information from disparate data points was untenable. Operators could selectively pick and choose what data to act on, resulting in an incomplete picture. Not to mention, with more widespread IIoT-enabled operations came an era of Big Data, and an ongoing struggle to make sense of the never ending flood of information.
Enter artificial intelligence and the power to ingest all the disparate bits and pieces of information, pull them together, and use AI’s processing prowess to make actionable insights.
As IoT has matured, Kathleen Mitford, corporate vice president of global industry marketing at Microsoft says it has become the foundation for AI in manufacturing. She says the adoption of AI is faster than anything Microsoft has ever seen. But AI isn’t a replacement for big data, Mitford says, but an enhancement — and multiple manufacturers are already putting it to use.
“The technologies that you’ve already been implementing over the last ten years in bits and pieces, like cloud infrastructure, managing data at the edge, and IoT, have come to a level of maturity enabling the pace of adoption for AI that I have not seen in my 30-plus year career working in and around the tech industry,” Mitford says.
AI in Product Development
Harting is one of the largest manufacturers of industrial connection technology. The company has a large catalog of connectors used in industrial applications. Despite its robust product portfolio, the company’s engineers perform a lot of custom design work. Now, Harting is using AI to make custom-designed products significantly faster.
Previously, when a new product request came in, a CAD engineer would search Harting’s library to build a custom CAD model, and design custom components. The process took hours.
The company recently partnered with Siemens and Microsoft’s Azure AI platform to use artificial intelligence to create custom designs. Now, the CAD engineer loads the project requirements and the AI scours Harting’s entire library to create a precise CAD design. The tool has not only freed up engineering department resources, but it does a better job of using up existing inventory.
AI is also being used for generative design work at Harting, capable of creating custom, more sustainable products and accelerating the design for manufacturing (DFM) process.
“AI is changing the design process. It’s being used to allow [companies] to create products that are more personalized and innovative, which allows them to respond to customer demand faster and increase brand loyalty,” Mitford adds.
AI is Running Factories
Bridgestone is a multinational tire manufacturer that has been in business for nearly 100 years. Still, the company runs into manufacturing downtime and bottlenecks. Previously, when operations hit a snag, engineers would chase down the root cause for hours. On average, it took about four hours to look through production data and system monitoring assets to find a fix.
The company partnered with Microsoft service provider Avanade to build a custom AI tool that uses Microsoft Copilot, the software giant’s digital AI assistant. Now, when bottlenecks happen at a Bridgestone facility, Copilot can not only flag the inefficiency, but come up with a solution.
For example, frontline workers can ask Copilot a question, like “Where is the production slowdown coming from?” Copilot searches all production and asset monitoring data and identifies the problem.
AI in Maintenance
Despite its status as a multi-billion aerospace company, Textron has struggled with downtime. Specifically, that associated with the time its airplanes were grounded. The company deployed Microsoft Copilot to not only identify the problem and provide step-by-step solutions, but also to pinpoint what talent (i.e. maintenance technician) was most qualified for the job.
Textron built a custom Copilot and fed it 60,000 pages of maintenance documentation and manuals. The smart assistant was up and running in just a few weeks, and has since been rolled out to all of Textron’s 1,600 technicians across 20 global facilities, serving more than 250,000 aircraft.
As veteran manufacturing maintenance talent has retired, the industry has run into problems as the “machine whisperers” left the plant floor without any tangible way to “download” their many years of expertise. But that tribal knowledge is recorded someplace, like those 60,000 pages of documentation. Now, Textron’s Copilot tool is even being used to more efficiently train new workers.
Manufacturers have been hesitant to adopt AI because it’s such a nebulous buzzword that many don’t know where to start. Mitford stresses the importance of looking for applications where you will get the biggest bang for your buck, like improving efficiency and keeping product lines running.
And workers are not as averse to using AI in the workplace as some may think. Like all other technology that has migrated into the workplace, when employees use a technology in their personal life, they expect the same level of sophistication on the job.
Where to Start
Like any expensive new tool you want to bring into the shop — especially one poised to be so transformational — Mitford says manufacturers need buy-in from the top, executive and board-level support. She says those in-house AI champions need to focus on how AI can be used to drive the business forward.
Most manufacturing operations already have some form of IoT infrastructure in place, so Mitford stresses that AI adoption isn’t as big of a lift as some anticipate. The time to hesitate is through. “AI is real and happening now,” Mitford says. “If you haven’t already started on the AI journey, you’re behind.”