Bridging the gap between algorithms and field exploration
Montero Mining and Exploration is laying the groundwork for the effective use of artificial intelligence (AI) and machine learning in mineral exploration by integrating data obtained from its exploration programs in Chile with these advanced technologies. The company is collaborating with US-based AI specialists to help define where and how AI can add value in early-stage exploration, using geological insight and advanced data science to identify potential mineral targets.
Dr Tony Harwood, CEO of Montero, says the company’s contribution to the development of AI for mineral exploration is both technical and practical, helping to bridge the gap between algorithm design and on-the-ground exploration.
“The data feedback from our projects is being used to refine AI model parameters and improve predictive accuracy for the wider mining sector,” he says. Montero follows an integrated data approach that brings field and regional data into unified datasets for analysis by AI and machine-learning models. These systems detect subtle patterns or anomalies linked to geological features and potential mineralization.
The company’s exploration team collects and validates high-quality field data through mapping, geochemical sampling, geophysical measurements, and remote sensing at its exploration projects. It sources large regional datasets from publicly available geological databases in Chile, historical records, and satellite data.
Harwood explains that machine analysis doesn’t replace the company’s geologists but significantly strengthens their ability to investigate beyond conventional boundaries and prioritize targets more efficiently. AI groups are collaborating with Montero because the company is open to new technology and offers extensive geological experience along with high-quality, structured datasets suitable for machine learning.
“We provide a real-world testing ground in Chile’s prospective mining belts,” says Harwood. “The collaboration helps refine AI algorithms in complex geological terrains where pattern recognition is particularly challenging. It is a two-way exchange: we bring exploration field data and context, and our AI partners bring advanced modelling and computing capabilities.”
AI models
Montero is testing a suite of complementary AI models designed to process different types of exploration data. “We have found that no single model can interpret geology, so we combine specialized algorithms that each contribute unique strengths to the predictive process,” says Harwood.
“We also test unsupervised clustering algorithms, which group data into meaningful geological domains without prior assumptions. These models reveal previously unrecognized relationships between geochemistry, geophysics, structural features, lithology and alteration.”
The integration of these AI models gives Montero a multi-dimensional predictive capability, linking surface geochemical patterns, key structural features in the rock, and spectral signals from satellite data into a coherent mineral-targeting framework.
The result, Harwood explains, is a more accurate, data-driven prioritization of exploration targets that helps guide the company’s field program and drilling strategy.
Selecting and adapting AI models
When it comes to selecting the appropriate machine-learning models, Montero starts by defining the geological question. For example, identifying porphyry-style alteration or detecting specific geochemical signatures. The AI teams then choose or adapt models suited to the data type, for instance, convolutional neural networks for image-based datasets and gradient-boosting algorithms for numerical geochemical data.
Model training and validation are carried out using known deposits as analogues. Montero’s geologists test the models’ predictions in the field, closing the feedback loop between machine output and real-world results.
Harwood says early AI applications have been most effective in detecting geochemical anomalies, unusual chemical signatures in the rock, and in mapping alteration zones from remote-sensing data. The models have been able to identify subtle spatial relationships between datasets, including faint geochemical trends and structural patterns that were missed through manual analysis.
“However, AI also introduces complexity,” says Harwood. “It requires rigorous data cleaning and strong geological validation to avoid false positives. The real gains are in speed and precision, generating target zones that can be verified in the field in days rather than weeks.”
Project impact
At the company’s Avispa copper‑molybdenum project in the Atacama, AI-driven modelling has accelerated interpretation by integrating surface geochemistry, regional structural features, and remote-sensing datasets into a single predictive framework. “It’s helping us prioritize drilling areas based on geological logic and machine-learning predictions,” says Harwood.
Meanwhile, at its Elvira project in the Maricunga mining belt, AI-driven data integration has highlighted previously overlooked alteration zones, which Montero has since confirmed through field sampling. “These advances demonstrate how multidisciplinary collaboration can turn raw data into meaningful discovery opportunities,” he adds.
Innovating with AI
Montero Mining’s innovation lies not only in its use of AI but in how it integrates diverse datasets into a single interpretive framework. Its geological team combines remote-sensing, geological, geochemical, and geophysical information to reveal patterns and relationships that remain hidden when each dataset is viewed in isolation.
Harwood says the real value comes from being able to test geological hypotheses rapidly through AI-driven pattern recognition and cross-validation. This capability allows Montero to generate, rank, and refine exploration targets at a pace that traditional methods cannot match, a shift that is reshaping expectations around early-stage discovery.
He notes that effective exploration with AI requires experience, discipline, and a mindset geared towards continuous learning. High-quality, well-structured data is essential, yet many mining-sector AI initiatives falter because datasets are inconsistent or incomplete. “You need the in-house geological expertise to curate and verify input data before it enters any model,” he says.
For Montero, which has long applied advanced exploration and metallurgical techniques, AI is not a departure from its approach but the next logical progression in strengthening decision-making.
“It helps us operate more efficiently, reduce exploration risk, and make better choices about where to invest time and capital,” says Harwood. “The aim is not to outsource discovery to technology but to use AI as a catalyst for faster, evidence-based exploration.”