Mining companies that are implementing advanced analytics need to fail fast, avoid shortcuts and prioritize value, according to Greg Brouwer, senior vice-president, technical at Teck Resources, who gave a keynote address and then answered questions on the theme “Data mining: mining operations in the digital age” at CIM Connect on May 15. 

Brouwer, who is responsible for Teck’s enterprise-wide innovation, transformation, technology and digital systems, talked about the role that technologies such as machine learning and artificial intelligence (AI) can play in helping miners adapt and thrive, but acknowledged “the frustrations many of us feel as we try to navigate this unfamiliar landscape.” 

“Technology is a topic Teck has spent a lot of time thinking about and pursuing,” he told the audience. “We recognized the potential transformative impact that advanced analytics, including AI and machine learning, held for the mining industry. We set out to explore the application of these technologies to our operations in a committed and systematic way over a four-year period, and what we learned during that period has proved invaluable to us.” 

Brouwer noted that the mining industry can be risk averse when it comes to new technology. “Given the nature of our work, we look for ways to incrementally improve,” he said. “But that means at times we can also be later adopters once the risks are well understood. As an outcome, the mining industry often appears at the bottom of the global R&D and innovation rankings compared to some other industries.” 

However, he pointed out that the concepts of advanced analytical advisory systems and machine learning have been around for a long time, and have been leveraged in other industries for many years before being picked up by the mining industry.  

“They are well proven technologies, but the problem is that we haven’t had the computing horsepower that’s required to crunch the numbers for some of these algorithms,” he said. “Another major issue is that as an industry, we didn’t necessarily have all the data that we needed, and we weren’t able to get it to the cloud fast enough to be able to do the things that we now do.”  

In Brouwer’s opinion, any modern-day mining company going forward needs to be looking for opportunities to deploy these technologies. “Because if you’re not, I think you’re leaving material value,” he said. 

He noted that the mining industry has only scratched the surface of technological innovation, which makes it easy to become overwhelmed. “When I find myself talking about technology to a roomful of miners, I can often see hesitation and skepticism,” said Brouwer. “Many of us have had bad experiences with technology that overpromises and underdelivers, that looks and sounds great but doesn’t actually have the impact we were sold. At Teck, we have had plenty of those experiences.” 

According to Brouwer, these negative experiences are what led Teck launching an ambitious program in 2019 to develop digital tools that use AI and advanced analytical software solutions but that were user friendly and reflected the complexity of mines and smelters. Five years later, Teck is Microsoft’s largest cloud storage customer in Western Canada, with over 850 terabytes of data stored in the cloud. “For years, we didn’t realize how valuable [that store of data] is, in large part because we didn’t have access to the incredible power that cloud computing now affords us. We didn’t have the expertise internally to unlock that value,” said Brouwer.  

“With AI and advanced analytical techniques like machine learning, we can sift through this wealth of information to uncover hidden patterns and insights.” 

These machine learning algorithms can identify trends and relationships that are difficult for humans to identify. “For example, they may notice that certain conditions in the mine lead to higher yields, or that specific equipment configurations result in more efficient operations,” he said.  

These algorithms can then be used as advisory systems to optimize an operation in real time. “In essence, these advisory systems become our digital partner because they work alongside us to continually improve and optimize every aspect of our operations where they are deployed,” said Brouwer. “We’re not just mining for minerals; we’re mining for insights, driving efficiency and unlocking value added return.” 

For example, Teck was able to use machine learning directly on its concentrators—which are often an operational bottleneck—to analyze and optimize grinding and flotation. “We built several AI-based advisory systems that [advise] on what changes we need to make in our concentrators in real time to maximize their performance,” said Brouwer.  

“You don’t always have sensors and data in all the places in the circuits that you’d ideally want, but that ended up not being a problem at all, because we were also able to use the exact same technologies and AI to develop something called soft sensors. This is technology that uses complex models to basically fill in the blanks when you don’t have physical sensors to give you the information you’re looking for.” 

Brouwer noted that Teck’s work on optimizing its concentrators has been some of the most valuable work that the company has done, and these tools unlocked several percentage points of metal production improvement.  

Brouwer added that it is important for everyone at the operation to understand what the core technologies around machine learning actually do and how they work. “One of the main things that we did provide training across all of our operations so that people understood these core concepts,” he said. “It wasn’t because we wanted everybody to become a data scientist, but we wanted them to understand what these tools and people like data scientists can do when you unleash them in our industry.” 

He noted that Teck had no difficulties hiring people with expertise in areas like data science to work in the mining industry. “In the space of AI, specifically, there’s just incredible opportunities to drive positive outcomes from a sustainability perspective,” he said. “The people that we ended up hiring were craving the opportunity to work in an environment like this.” 

The digital innovation playbook  

While Teck has had great success implementing these technologies, Brouwer noted that the company has also learned some difficult lessons along the way. “Innovation on the scale that we’re driving isn’t easy, and we made mistakes,” he said. “But we learned from those mistakes, and the way we overcame them helped us to create our own playbook for fast-tracking digital innovation within our organization today.” 

Brouwer shared the three most valuable lessons that Teck has learned, the first of which is to fail fast. “That is actually a core tenet of agile project management and product development,” he said. “You need to be disciplined about putting those products that don’t return value on hold or cancelled as quickly as you can. Because it’s once you have the proof of concept that the real work begins—80 per cent of the effort is expended on refining that concept—so don’t get bogged down in perfecting those early proof of concepts or models. Find the core value early, and then you can worry about adding more scope and more functionality later.” 

The second lesson is to avoid shortcuts. “Taking shortcuts on technical processes and when putting the application together can significantly impact your ability in the long run to support and sustain the applications in a cost-effective way,” said Brouwer. “Those applications, if not built properly, can cost you a lot more than you think down the track. This is something that we learned a particularly hard lesson around.” 

To avoid this, Teck invested in building a platform that was specifically designed for the company to deploy AI-based tools at scale. “Now that this platform is in place, we can automate and accelerate product innovation and deployment, so we can get products and product enhancements into the hands of our mines sooner,” explained Brouwer. “When we need to scale up to support more of our main operations, we can respond faster and at a far lower cost. Going forward, everything we do will be faster, more stable, more secure and more sustainable as a result.” 

The third lesson is to prioritize value. “Throughout the entirety of the innovation life cycle, we measure the impact and value of the technologies we deploy,” he said. “It’s tempting to skip this part of the process because it can be very difficult and time consuming to try to isolate the impact you’re having on very complex processes with these digital tools, and the development of the technology itself can be challenging enough.” 

However, Brouwer thinks this was a key differentiator in Teck’s approach compared to some other companies, and guided the company to make better decisions. “If [a technology] moves things forward, then we proceeded,” he said. “If it couldn’t, we retired it, no matter how cool or innovative it seemed.” 

These simple tenets helped Teck to stay on course and unlock real, tangible and measurable value from technology, said Brouwer, adding that the company is now moving to embrace tools like generative AI and other “new and exciting” technologies that are more inclusive, accessible and dynamic.  

Brouwer thinks that generative AI will be a transformative technology, not just in mining but across society in general, but it will not replace traditional AI. “I’m really excited to see how this technology that we’re just starting to scratch the surface of right now is going to profoundly impact our industry and the way we interact with technology in the future,” he said. 

“I believe it is inefficient if each of us works on these technologies in complete isolation. The future that I’m hoping for is one where we come together as an industry to share insights, our progress, our lessons learned, and use them to find solutions and amplify our collective impact.”