Episode 1: A Front-Row Seat to Mining Evolution | Ft. Kirk Petroski, Founder of SYMX.AI
The mining industry is undergoing a significant transformation as operations increasingly turn to data, automation and AI to improve productivity, safety, and decision-making. In this conversation, we sit down with Kirk, founder of SYMX.AI to discuss the challenges facing modern mining operations, the role of operational intelligence in driving efficiency and how AI is helping mines unlock value from the vast amounts of data generated every day.
1. You’ve spent over three decades in the industry, how would you compare the state of technology in the 1990s to what we see today in the 2020s in the mining industry?
Kirk: Biggest evolution has been the availability of digital wireless connectivity and digital data that had never existed prior today, both are critical for immediate and long-term decision-making. When I started my technology journey in the 1990s, the biggest advancements for industrial use was satellite, GPS-based location data for mine exploration. Another one were sensor devices such as magnetometer survey gear for earth penetrating that manually logged the data. Connectivity then was via a serial cable (RS-232) to the first Panasonic Tough Book for collecting geophysics data. We also logged GPS location data manually.
With the use of a laptop and software I was able to synthesize the data sources and visualize and report the data on an exploration survey map. That was ground-breaking back then! Connectivity and data analysis at the frontier level was very slow but a great starting point to build from. Today, with Wi-Fi, Public/Private LTE/5G and satellite, connectivity has become almost pervasive, and we can generate amazing amounts of data.
2. Everyone talks about AI transforming the mining industry, but the industry still struggles with fragmented and inconsistent data. As per you, how ready is the industry, realistically, to support large-scale AI adoption from a data infrastructure standpoint?
Kirk: AI adoption today is reminiscent of the connectivity and data availability frontier of the 1990s. We are learning as we go, but importantly we are willing to move the industry forward. Yes, fragmented and inconsistent data persist but that is the starting point. As we improve with data structure, industry standardization and better interfacing methods, we will be training AI on proprietary data sets that are relevant to our mining industry needs. The infrastructure of today that enables data processing and AI in the Cloud is also a key infrastructure point. Previous on-premise data storage and compute had scalability limitations.
It is great to see the mining industry adopt cloud storage and compute necessary to run centralized AI. We will see a continuous evolution here with hybrid cloud/prem systems where localized and decentralized AI can grow. The early wins need to be amplified and I am very excited about the work that SYMX.AI is doing on this new frontier for the mining industry.
3. Where do you see AI delivering the most immediate, practical value across the mining lifecycle – exploration, operations, processing or somewhere else? And where do you see the biggest friction even today?
Kirk: If we start from data availability and consistency framework then processing would make the most sense to start with. This is because fixed assets for processing are already connected and can deliver consistent data over a physical Ethernet network. This is a continuous process. When it comes to the drill, blast, load, haul and dump cycle, this can often be considered batch processing and subject to inconsistent data, connectivity and workflow challenges. This is a new frontier for AI adoption with tremendous upside for the mining industry. However, for the mining industry to support the macro level global changes such as AI data centres, electrification and defence, accelerating discovery of new ore bodies, expanding orebody knowledge and building mines quicker is the most critical.
Remember technology is only a 1/3rd of adoption. People and processes have such a large role in adopting AI. Understanding of where AI can support within existing workflows, then improve workflow as this evolves is of utmost importance. People’s role in adopting AI is changing as individuals adopt AI in their personal lives and then see the benefits it can bring to their operations.
Early in my career, one of the biggest challenges was establishing trust in data, and I believe we are at a similar stage with AI today. The difference is that instead of spending our time analyzing raw data, we are now evaluating whether AI-generated insights are accurate, reliable, and actionable enough to support decision-making. Ultimately, AI adoption is built on trust. Organizations will need to make results visible, demonstrate successes, and benchmark outcomes. As confidence grows through proven performance and validated outcomes, trust in AI will follow, enabling broader adoption and greater operational impact.
4. Do you think the industry has the right mix of talent to adopt AI at scale, or is the skills gap a bigger constraint than technology itself? Is that the main reason why many AI initiatives don’t move beyond the pilot stage?
Kirk: We are an industry of fragmented specialists with the desire to work together collaboratively to advance our mining industry in the areas of safety, sustainability and efficiency. This has been the design for progress in areas within the mining lifecycle. I have no doubt that the education, background and experience of mining professionals have allowed the industry to be successful at a given rate and scale. However as demand grows, resources become more remote and deeper. And with fewer people entering the mining industry, it will take a massive effort to scale and this is where I see the role of AI.
Just recently, OpenAI announced that one of its “general” reasoning models solved a 80 year unsolvable math problem (Erdos Problem #90 – Planar Unit Distance Problem) in just over 3 hours. Put that into perspective and imagine what AI can do with available mining data and information. The mining industry has its specialists with knowledge of specialized data. Imagine the endless possibilities if specialists enable AI on top of this data. AI doesn’t have multitasking, reactive situations, lunch, coffee breaks, vacation, etc. Give AI the task, validate the findings, and if they are accurate, act on them. Focus on incremental wins rather than disruptive step changes. That is how the industry moves beyond the pilot stage and delivers results that matter.
5. As critical mineral discovery becomes central to energy security, do you see AI-driven resource intelligence becoming a geopolitical advantage & not just a commercial one?
Kirk: 100%. This has been a hot topic since governments worldwide prioritize critical mineral discovery and strengthen efforts to secure strategic mineral supply chains. A great read on this topic is “Mining is Dead. Long Live Geopolitical Mining” by Rivera and Zamanillo. The world has finally woken up to mining as a strategic priority as Nations with critical minerals will have strategic power and be major players in the next global order. At the same time, governments are developing their AI and IP strategy globally. Mining jurisdictions such as Canada can have a critical role to play in advancing AI that supports the mining industry.
6. If we fast-forward to 2036, what will have changed more: the technology itself, the regulatory environment, or the way mining companies fundamentally operate? What does the future look like?
Kirk: Building on the foundational layers of pervasive connectivity, massive data availability across the mining lifecycle and AI driven intelligence, the next frontier will be the convergence between the physical and AI world. Today, we can use AI as a tool to support decision making and automate some of the process. By 2036, mining operations will achieve interoperable autonomy, where physical machines and human workers seamlessly communicate, adapt, and coordinate with one another, regardless of brand, platform, or software.
Over the next decade, with advances in embedding intelligence, with the work performed (by people and machines) and a data layer captured from existing workflows will set the stage for what’s to come. Mining companies will increasingly treat the data, information, and knowledge generated by their workforce and equipment as valuable intellectual property. This will create a foundation on which AI can continuously learn and build.
We are seeing autonomous mobile equipment, drones and human hand held technology, but it will be a matter of time before workers are supported by front-line humanoid or other robots. One who will be capable of operating existing non-autonomous equipment, while being connected to the mine plan (annual, monthly, daily tasks), mine safety regulations, SOPs and other foundational data.