Move over, robots. Large language models (LLMs) may soon be joining you on the factory floor.
Microsoft Corp., OpenAI and Alphabet Inc.’s Google are sending chatbot-like tools to industrial facilities as AI unlocks manufacturing processes. Recent breakthroughs in LLMs in particular have proven potential money-makers as companies look to cut down on defects, downtime and wasteful energy consumption.
German manufacturer, Schaeffler Technologies, offers a compelling example of what can be achieved. It is modernizing data management across diverse IT/OT systems with generative artificial intelligence (AI), leading to advanced factory reporting and troubleshooting to improve uptime, quality, productivity and yields.
Last year, Schaeffler became one of the first users of Microsoft’s Factory Operations Agent, a new product fueled by LLMs designed for manufacturers. The agent is scheduled to leave public preview later this year, when it will be widely available to Azure AI users.
The agent, deeply integrated into Microsoft’s enterprise products such as data-analytics system Microsoft Fabric, lets Schaeffler operate data from hundreds of plants globally. That is especially crucial where tracking down errors in manufacturing means comparing data across quality assurance systems, HR software and industrial control systems.
Schaeffler is among the highest-profile examples of work Microsoft is doing with manufacturers.
Rockwell Automation is providing industrial AI expertise via the Azure AI model catalog for recommendations, explanations and knowledge about specific manufacturing processes, machines and inputs to factory floor workers and engineers. Meanwhile, Sight Machine, which specializes in data-driven manufacturing and industrial AI, will release Factory Namespace Manager to the Azure AI model catalog. The model analyzes existing factory data, learns the patterns and rules behind the naming conventions, and then automatically translates these data field names into standardized corporate formats.
Microsoft is among a growing list of tech companies deploying AI on the factory floor for predictive maintenance, quality control via computer vision, and robotics and cobotics to deliver clear and immediate ROI, according to Relay CEO Chris Chuang.
[With more than 750,000 robots at its manufacturing operations, Amazon.com Inc. is among the largest employers of machines, but it does not disclose the size of its warehouse workforce. The company, in total, has 1.55 million employees worldwide, many of them in fulfillment centers.]
LLMs like the Factory Operations Agent address a so-called IT/OT gap within the industry, connecting spreadsheets and other tech information with operational tech used in a factory.
Verkada’s customers are leveraging its advanced AI-powered analytics and features to detect and alert teams of safety risks in real time, such as a forklift operator wearing headphones or employees without personal protective equipment.
“Other customers are using it for quality control on the assembly line. For example, an auto manufacturer is linking video footage with assembly line data to enable them to instantly pinpoint defects like chips or dents by searching for VIN numbers, helping teams quickly identify and resolve issues with video verification,” Abraham Alvarez, vice president of product management, cameras, at Verkada, said in an email.
Late last year, Google updated its Manufacturing Data Engine to unlock data on industrial devices. Microsoft and Google also provide platforms where independent developers can test systems, based on their strategies and risk tolerance. In fact, safety remains a primary concern in the race to industrial AI, especially on the factory floor where malfunctions can lead to serious injury, including death.
