Caterpillar is transposing many years of mining automation experience into broader AI applications. The management of the company stated that the real difficulty was not just to develop technology, but to allow AI to access the process on the ground and to work with existing personnel.
From mines to sites
Caterpillar was the first to advance automation, mainly from the mining landscape. Labour shortages and high-risk operating environments make mines more valuable for automation. Today, the company has sold self-piloting trucks, rigs, sub-surface loaders, bulldozers and remote control construction equipment, with software tools such as command centres, fleet management and terrain intelligence.
Jaime Mineart, the company ' s chief technical officer, stated that Caterpillar now wished to bring these experiences to a more dynamic environment, including construction sites, quarries and construction sites. She argued that the deployment of automation equipment did not amount to the completion of the AI conversion on site, and that the enterprise needed to re-engineer its operating processes and reorganize the way people interact with machines.
AI Assistant into maintenance
In specific applications, Caterpillar has introduced AI into the tools used by technicians and internal employees. Its Cat AI Assistant supports on-site technicians to access maintenance processes through voice, to check for potential malfunctions and to identify spare parts that may be required prior to maintenance.
Mineart states that this tool is currently being used by clients, equipment operators and technicians. The assistant operates with his own data from his network system. She stated that the company owned about 1.6 million networked assets worldwide and accumulated more than 16 PBs of structured data.
Five years of training in $100 million
In addition to maintenance, Caterpillar is using AI to support on-site scanning and digital twin production to analyse manufacturing operations. It also uses AI for in-house operation and software development, including the adaptation of old codes, generation and testing of new software with AI agents, and early identification of deficiencies.
However, the company believes that the difficulty of AI landing remains at the organizational level. Mineart states that Caterpillar will rely on experienced operators to train AI systems to translate the in situ knowledge accumulated over the years into modelling capabilities. With the increased automation of equipment, some of the operators may move in the future from controlling single-station equipment to managing multiple equipment simultaneously in a remote command centre.
This shift has also created new training needs. Mineart states that Caterpillar plans to invest $100 million over the next five years to train its 118,000 employees in AI, automation and robot-related skills.
At the operational level, this input is also related to the rising infrastructure needs of AI. Caterpillar received $20.5 billion in the second quarter, a record high. Of these, sales increased by 72 per cent to $3.1 billion, driven by demand for power generation equipment in the data centres. Joe Creed, CEO of the company, stated that the demand around cloud computing and generating AI infrastructure continued.