For most people, the primary purpose of drill targets is to find the locations that produce the biggest results possible in terms of grade and interval length. While drilling monster holes full of wide bands of gold, silver, or whatever other mineral that is being chased is always welcome, there is another lesser-appreciated purpose to drill targeting, particularly in the domain of AI/ML targeting. Recognizing and understanding this other objective helps to identify hidden value in exploration companies - especially within the context of greenfield exploration.

One of the reasons that the application of AI and Machine Learning to mining exploration has attracted so much attention is the potential to examine massive quantities of data and to find and interpret slight variations and patterns that wouldn’t otherwise be seen by an individual. Unfortunately, in the early days that AI/ML was first being rolled out within the mining and geological community, it might have suffered from some irrational exuberance with some of the initial proponents of the technology suggesting that AI/ML was eventually going to remove the need and utility of the human geologist from the exploration process. On the opposite side of the spectrum were a number of feet-in-the-mud old schoolers that completely dismissed the technology out of hand as nothing more than smoke and mirrors - “I don’t need some machine-thingy to tell me how to do my job...”  The reality (as with most things) is that the truth lands somewhere in between these two extreme positions.

Effective utilization of AI/ML doesn’t replace the human geologist, but it does have the capacity to equip a geologist with better data and to improve their capacity to identify and interpret even subtle patterns and probability changes. Similar to how a person doesn’t need a set of binoculars in order to see, using binoculars when looking over a long distance enhances the visual capacity without obviating the use of the eyes. This enhancement of a geologist's analyzing capacity is particularly pronounced when the AI/ML is able to integrate and overlay data derived from different geological disciplines (i.e. structural geology, geophysics, geochem, etc.) into a single interpretive model. It is their specialization of this multi-disciplinary approach that is one the major reasons I am such a fervent proponent of Goldspot Discoveries ($SPOT) and their methodology of applying their AI/ML tools into mineral exploration. Most competent geological professionals can do a geochem analysis or interpret the results from a geophysical survey, but the special sauce in this domain comes when AI/ML technology facilitates the interpreting geologist gaining the ability to analyze the results from both of these disciplines (along with numerous others) contemporaneously with each other.

Another underrated capacity of AI/ML is its ability to develop probability models for the unknown structure under the ground. A good analogy is to think of the geologic modeling as a game of Sudoku. In the early parts of game, when very few of the squares are known, each additional square in which the answer is determined not only provides the result for that specific box, but also provides important information that facilitates the determination of the values of several other boxes.

If the property being explored is a site that has already been extensively drilled and there is a substantial amount of information known, there isn’t as much value in using AI/ML in interpreting the larger overall structure, as it is already known. In that scenario, what is needed is pinpointing the granular details of the mineral deposit as well as determining how to get the drill locations that maximize results. To use a different analogy, if you already have a detailed map (extensive drilling and other site data), the most important objective is to find an exact location on that map. But, if there is little to no geographical data known at the outset, it is incumbent to first develop a basic structure for the map. In that second context, even if a person was provided an exact address, that information has limited utility without knowing where the streets, highways, rivers, bridges and other essential components are in order to plot a course to the intended destination.

Going back to the Sudoku analogy, what is the most valuable target to drill on a greenfield project where little is known? The best target is the one that provides the most assistance in interpreting the remainder of the puzzle. And so it is with geological targeting. Sometimes the best targets are the ones where the results tell you the most about the underlying geological structure. Now some might say, “Who cares? Big drill results make the stock go up and that is all I care about.” That is a fair point if all you are looking for is a short-term pump.  But even an initial eye-popping result will lose its luster if the junior is not able to follow it up with additional results. It is a fool's errand to expect to produce strong repeatable drill results without a solid understanding of the underlying geologic structure. Long-term exploration success requires avoiding the temptation to try to put the cart before the horse.

Sterling Metals - Case Study on AI/ML Value-Add in Greenfield Exploration

A great example of how AI/ML can play a valuable role in greenfield exploration (where very little data is known) is Sterling Metal’s Sail Pond Project. Sail Pond is located in the Great Northern Peninsula of Newfoundland (not far from the earliest known Viking settlements in North America). While there is a considerable amount of gold exploration activity currently in Newfoundland, Sail Pond is quite geographically removed from the action that is generally clustered along the Valentine Lake and Appleton Faults. 

 In addition to being geographically removed from the area of the Newfoundland gold rush, Sail Pond is distinct from most of the other exploration on the island as it is a silver/base metals (copper/zinc/lead/antimony) prospect.

Sterling Metals obtained Sail Pond from Altius Minerals ($ALS) in October 2020. Altius received a 13% equity interest in Sterling as well as retained a 2% NSR on Sail Pond. The property has a bit of interesting history associated with it. The prospect was originally discovered in 2016 by local school principal and geological hobbyist, Tony Kearney, who tested some samples from an outcrop on the property. It was Kearney who initially staked the claims and then subsequently sold the property to Altius Minerals. The first company to option the property from Altius was actually New Found Gold ($NFG) back in 2018. After New Found Gold optioned Sail Pond, they spent $1.5 million conducting an IP survey of the entire property. Although the Sail Pond prospect and IP survey results were very promising, New Found Gold allowed Sail Pond to revert back to Altius as the stock-heavy deal for Sail Pond was no longer economically sensible after the discovery of gold on its Queensway property, which also caused New Found Gold’s stock to balloon 20x in short order.

Immediately after New Found Gold surrendered the property back to Altius, Sterling Metals (who was called Latin American Minerals at the time) jumped in and swiped up the property in October 2020. It is worth noting that the CEO of Sterling Metals, Mat Wilson, has past professional ties with Denis Laviolette who is the Co-Founder and President of New Found Gold as well as the Executive Chair and President of Goldspot Discoveries. Both Wilson and Laviolette worked together in a prior life at Pinetree Capital. Laviolette also serves as an advisor to Sterling Metals. From the connections between the parties and the timing of New Found Gold’s surrendering Sail Pond and Sterling Metals’ moving in immediately after to swipe it up, it is reasonable to assume that there was a very smooth transfer of all the exploration data obtained by New Found Gold to Sterling Metals.

One of the first things to notice about Sail Pond is the size and scale of the project. The project runs in a southwest to northeast direction over a distance of more than 12 km. 

 Prior to optioning Sail Pond to Sterling, Altius did do some consolidation of the Sail Pond claims so that the current claim boundaries only cover the major mineralized zones. This isn’t a project with a lot of surrounding land that is held for no particular reason other than it is adjacent to the property. This is good for Sterling as it helps avoid incurring unnecessary cash burn associated with maintaining claims that hold little long-term economic value. Even while the mineralization appears to run in a tight band structure, the main mineralized zone that is being actively explored is nearly 25 sq. km. This is an immensely large exploration property.  A cool graphic produced by Sterling that is very helpful in providing context for the size of their project was an overlay of Sail Pond over Manhattan Island.  

To make the challenge more daunting, as recently as a year ago, no drilling had ever been performed on the property and very little was known about the underlying geology. In fact, there is very little prior exploration work throughout this entire part of the Great Northern Peninsula from which Sterling could extrapolate prospective geological structure. This project puts the green in greenfield.

This is where the power of AI/ML-enhanced geology comes into the story. Although Sterling holds an interest in an immensely large set of claims, Sterling’s market cap is a modest $17mil. While its drill programs have been fully funded, Sterling is heavily incentivized to maximize the bang for its buck with the exploration dollars it does spend. This is easier said than done on a property that has never been drilled before. Granted, it was a nice little house warming gift and head start boost that Sterling was handed the results of a $1.5M IP survey essentially for free. But regardless, exploration cost-efficiency is crucial in this context.

Sterling Metals has a close relationship with Goldspot Discoveries and Sterling retained Goldspot from the outset to assist with its exploration of Sail Pond. Sterling’s initial drill program was a modest size of 7,500m done over 4 months using a single drill.

After only the first few drill holes, Sterling (with the assistance of Goldspot) was able to identify a contact point between the host dolostone and an argillite footwall that appeared to be strongly correlated to higher mineralization. Sterling was then able to use the results from the prior IP survey done by New Found Gold to identify and map the entire dolostone/argillite contact point across the entire 12km length of the project. 

After doing this, Sterling was able to reorient its remaining drill holes mid-program to begin focusing on the immediate proximity of the newly mapped contact zone. In late-February, Sterling released the remainder of the results from the inaugural drill program, including the announcement of the discovery of the Heimdall Zone.

In a shareholder webinar in late-February, Kelly Malcomb, Principal Geologist for Sterling, demonstrated 3D modelling of the newly discovered zone. <A replay of the webinar can be found here:>. While the Heimdall Zone is a nice discovery in its own right, one of the most remarkable things about Sterling’s exploration of Sail Pond was how quickly its geological team (again with Goldspot’s assistance) was able to identify the underlying structure driving the system and be able to specifically target the dolostone/argillite contact area after only a few drill holes.  Below is a screenshot from the webinar showing the currently defined boundaries of the Heimdall Zone:

 It is important to keep in mind how unique and important it is to have the capacity to almost instantaneously adjust the overall model and plan based on a few drill holes. Traditionally, an explorer working a virgin piece of ground would complete the drill program and then analyze the results during the down season and build in the adjustments for the following season’s drill program. The ability to make mid-program adjustments on the fly based on the results of just a few drill holes (aided by AI/ML modeling and adjustments) allows a junior like Sterling to shave years and millions of dollars off its exploration budget.

I have no doubt that provided ample time and budget, the Heimdall Zone would have been discovered through traditional geological methodologies. It is not making the discovery that demonstrates the value of AI/ML in greenfield exploration, rather the speed and adaptability that AI/ML provides. Using only the results of the first five drill holes, Sterling was able to completely modify its structural model, reorient its drills, and make a new discovery. That level of speedy adaptability saved Sterling millions of dollars that would have otherwise have been spent using traditional geological tools and resources. Now, heading into its second drilling campaign, Sterling Metals is well positioned to take the next step at Sail Pond. Sterling recently announced the commencement of a new 7,500m drill program. They are armed with 38 newly-identified priority targets. I am looking for two major objectives that should be accomplished with this year’s drill program: (1) significant expansion of the dimensions of the Heimdall Zone beyond its current identified size of 400mx200mx80m as the zone is still open in multiple directions; and (2) specifically targeting locations within the Heimdall Zone that provide some eye-popping intervals that demand attention from the market.

While Sail Pond is still an early stage greenfield project, there are lots of reasons to be optimistic about the future. Mat Wilson, Kelly Malcomb and the rest of the Sterling Metals team were able to make a solid discovery of a new zone in the very first season that a drill was ever put in the ground on the property. Armed with the knowledge obtained from last season’s drilling and utilizing some of the most cutting-edge AI/ML methodologies available, I am optimistic we will be seeing some fireworks near the end of the year as drill results become available.

Disclosure: I have a beneficial long position in the shares of one or more of the companies discussed in this article, either through stock ownership, options, or other derivatives. I wrote this article without external assistance, and it expresses my personal opinions. I was not compensated for this article, and I have no business relationship with any company whose stock is mentioned in this article.