Freeport-McMoRan: Transforming AI in mining | McKinsey Digital

McKinsey helped Freeport create a digital roadmap using AI, advanced analytics, and agile work methods to increase productivity at every step of Freeport’s processes. The new approach favored minimum viable products that could be continuously improved, rather than the traditional goal of “perfecting” a solution before it was deployed. To entrench this capability, McKinsey brought in agile coaches to train teams to operate faster and better.

“We put in the recommended AI engine and saw 10% improvement in production. And we thought if we do the implementation at all seven of our sites right, it's almost like having a brand-new plant without having to go through permitting process and disturbing a new area. It's in the billions of dollars that we're offsetting by going through the transformation.”
—Cory Stevens, President, Mining Services

Freeport had a big leg up for its AI transformation because it had built a central cloud-based data architecture. A key component was a data warehouse to store the data collected from sensors installed on the company’s trucks, shovels and stationary machines, allowing Freeport to capture second-by-second performance readings. It used that data to train an AI model custom-designed and built by McKinsey to find operational improvements that could increase output at lower cost. Instead of running the plant at a single setting all day, Freeport could now adjust settings every hour to maximize production from a given type of ore, quickly boosting production by 5-10%. The amount of additional copper production Freeport is projected to unlock over five years is equivalent to one new processing facility without the eight to ten year wait to bring a new facility online. Once the AI models were built in a modular way, Freeport was able to easily adapt and scale throughout its mines in the Americas.