Samuel Addo-Frempong, the developer of minexTRM platform and Mining Director at Tex-Mining, explains how an artificial intelligence (AI)‑enabled, economics‑aware target ranking model blends geology with operating and financial inputs to cut discovery time, sharpen drilling budgets, reduce land disturbance and give boards confidence across commodities, including critical minerals.
Tex-Mining won the Innovation award in the 2025 Mining Technology Excellence Awards for the AI-driven platform.
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
Mining Technology (MT): Samuel, congratulations on the Innovation award—when you look back to 2015, what problem in exploration decision-making compelled you to build minexTRM, and how has that mission evolved?
Mineral exploration is an expensive, time-consuming and risky adventure, since there is no certainty of significant find. This is what motivated me to develop a model which could guide companies towards channelling resources to targets with high potential to become major discoveries. Over the years, with the growing demand for mineral resources and the need to replenish depleted resources, it was evident that time was of the essence and there was the need to discover the next massive deposits faster. This was when artificial intelligence was incorporated into the model, to enhance data analytics and improve the accuracy of predictions, with the ultimate aim of reducing mineral discovery time.
MT: How did that original vision shape the product choices you made—what did you deliberately include (or exclude) to keep the tool practical for field and corporate teams?
Beyond geological signature, this model incorporates engineering, operational and financial information as model inputs. The idea is to assess potential operating parameters and economics of various exploration targets even at early-stage exploration. It is worth noting that outside geological context, operational and financial data have significant impact on the economics of a mineral prospect. This is the main driver for incorporating them into the model.
MT: Winning this award inevitably raises expectations. What does the recognition mean for your client relationships and your own goals as a leader?
Winning the innovation award demonstrates the positive impact this AI-driven mineral exploration target ranking model is making in the mining industry. We understand the expectations and are really looking forward to working with our clients on their various mineral projects. My personal goal is to address the pressing challenge our industry is currently faced with, which is to replace rapidly depleting mineral resources by leveraging on the benefits of this model in achieving this goal.
MT: Many tools promise smarter targeting. In day-to-day use, what changes first for a client—how they plan drilling, structure studies, or allocate budgets?
What sets this tool apart is the incorporation of operating and financial data, apart from geology. With the utilization of this tool, companies are confident in their ability to strike major mineral deposits when they commence advance drilling. Drilling budgets are prepared while focusing on promising exploration targets – saving cost, time, and the environment since this targeted drilling approach results in less environmental disturbance.
MT: Adoption is rarely just a technology issue. What were the biggest hurdles you encountered—data quality, culture, or workflows—and how did you help teams get past them?
The mining industry is relatively slow in adopting new technologies, probably due to a mixture of culture and quality concerns. What has helped with the adoption of this tool is its successful piloting phase. To convince clients about the effectiveness of this tool in ranking exploration targets, pilot projects are run using their known preliminary geological datasets. Once the tool ranks the various prospects and advises on which prospects to follow up with advance exploration in preparation for mine development, the clients become convinced when predictions align with the results of their advance exploration efforts.
MT: Trust is earned in the pit and the boardroom. How do you validate model outputs with geologists and finance teams so both feel confident acting on the rankings?
One approach that aided in model output acceptance was model results in alignment with mine site advance drilling results. Essentially, the model is used to rank and predict promising exploration targets based on the mine’s existing early-stage exploration datasets. Although results from advance / confidence drilling activities are available, early-stage datasets are deliberately used. The model outcome is then compared to the available results of the confidence drilling, and as they align, management teams get the conviction that the model works.
MT: One strength of minexTRM is comparing unlike targets on an economic basis. Can you share how that has influenced portfolio-level decisions without revisiting examples already in the public domain?
Geology alone is not enough in determining an economic mineral deposit. On top of geology, one needs to consider engineering and operating parameters, costs and commodity price forecasts. Putting all these together may alter the overall economics of a mineral prospect. In my opinion, a promising mineral target is one that satisfies all parameters, and these are the kind of discussions we engage with clients when developing their mineral exploration strategies.
MT: You operate across geologies and commodities, including critical minerals. What practical lessons have you learned about adapting the approach to different deposit styles and regions?
This is an interesting question. The nature and style of mineralization may be similar for certain types of deposits but not the same, as there is some uniqueness associated with each mineral deposit formation. In simple terms, the AI algorithms in the model are generated to identify these unique features and indicators to aid in pattern analytics as part of the prediction process. It is important to know the possible mineralization controls even at early-stage exploration to provide that background knowledge of potential geological signature.
MT: AI evolves quickly. How do you refresh models while keeping the user experience stable for teams in the field?
The AI algorithms work in the background during geoscientific data analytics, updates are generally independent of the user interface. Users can seamlessly use the tool irrespective of algorithm upgrades.
MT: Where do you intentionally keep a “human in the loop,” and how do you balance automated ranking with expert judgment at key decision gates?
Although the tool is essentially automated, human inputs are required in the areas of early-stage geoscientific datasets input such as various forms of geophysics and geochemical anomalies. Additionally, humans are needed to come up with cost inputs, potential operating parameters, and financial inputs such as commodity price and discount rates.
MT: Scenario and sensitivity analysis are central to the platform. How are clients using these tools to navigate price volatility, cost inflation, or metallurgical uncertainty in real time?
Typically, the tool makes it possible for clients to perform various sensitivities on the base case ranking. During this sensitivity analysis, clients can gauge how sensitive ranked prospects are to various operating and economic parameters. This aids clients in deciding which prospects to advance for eventual development based on how robust they are, under varying operating and economic conditions.
MT: Beyond discovery cost and reserve additions, which metrics best capture value over a multi-year exploration program—speed to decision, resource conversion, or something else?
The tool delivers value in four key areas – reduction in discovery cost, reduction in discovery time, reduction in land disturbance, and accelerated decision making—in other words accelerated definition of exploration strategy.
MT: ESG and efficiency increasingly go hand in hand. How does the platform help reduce unnecessary drilling, shorten campaigns, or lower land disturbance?
Imagine having several exploration targets and not knowing which of them to advance for development. You might end up drilling all of them to make that decision. In doing so, a large surface area will be disturbed. Whereas using this model shall narrow those targets to those promising ones for advance drilling activities, thereby reducing the land impact.
MT: Looking ahead, what capabilities are you prioritizing next and how do they advance your vision for objective, bias-aware exploration?
The demand for critical minerals is on the rise as they are essential for modern technologies, the economy, and national security. We see countries racing to dominate the critical minerals space. Going forward, more focus shall be given to identifying major sources of critical minerals. This objective makes it utmost priority to use this model to aid in locating major sources of critical minerals such as lithium, cobalt, copper, nickel, and rare earth elements.
About Samuel Addo-Frempong
Samuel Addo-Frempong is a mining engineer with over 16 years of experience in base and precious metals. He currently serves as Mining Director at Tex-Mining, a mining consulting firm headquartered in Dubai, United Arabs Emirates and delivering mining consulting services to clients globally. Samuel graduated in mining engineering from University of Mines and Technology in Ghana. He subsequently graduated (Summa Cum Laude) from the Colorado School of Mines where he earned his master’s degree in mining engineering and management.
Samuel’s area of expertise includes mineral exploration target generation, strategy development for mineral exploration, mineral resources and reserves reporting, mine construction and development, mine process optimization, M&A, technical studies (preliminary economic assessment, prefeasibility studies, feasibility studies), and mining technology and innovation. He has worked on mining projects and operations in Africa, Middle East, Latin America, and Eastern Europe focusing on gold, copper, cobalt, rare earth elements, lithium, nickel, and graphite.
Samuel is a Competent / Qualified Person as defined by Canada’s NI 43-101, U.S. SEC, Australia’s JORC, and South Africa’s SAIMM, and is a member of the Society for Mining, Metallurgy and Exploration (SME), American Exploration and Mining Association, and the Australasian Institution of Mining and Metallurgy (AusIMM).
Contact: samuel.addofrempong@tex-mining.com
