Applying AI to support better
mining decisions
De-blackbox AI through geological insights

Stratum AI Geospatial Estimator (SAIGE) creates a more accurate resource model with the data you already have

The Stratum Resource Model

Explore an AI-identified mineralization cluster.

The Impact of SAIGE on Model Performance

Drilling Efficiency & Resource Growth

Traditional Models
Static targets from averaged continuity
Low ounces per drill metre
Slow feedback into models
SAIGE - (DATA DRIVEN MODELS)
+32% ounces per drill metre
7.7 kt Cu identified in former waste blocks
16–24% drilling savings via confidence spacing
Observed in:
Orogenic Lode
IOCG
Intrusion-Related Reef Veins

Resource & Grade Modelling

Traditional Models
Manual domain smoothing
High block-scale uncertainty
Frequent waste misclassification
SAIGE - (DATA DRIVEN MODELS)
+43–56% prediction accuracy
Higher block-scale confidence
Up to −73% misclassified waste
Observed in:
Copper Porphyry
IOCG
Gold

Short-Term Planning & Reconciliation

Traditional Models
Static monthly planning assumptions
Plan vs actual production gaps
Frequent re-handling & re-planning
SAIGE - (DATA DRIVEN MODELS)
+2–6% higher mined grade
3–11% lower strip ratio
32–55% lower deviation in reconciliation
Observed in:
Copper porphyry operations

Processing, Recovery & Throughput

Traditional Models
Limited block-scale recovery prediction
Conservative mill routing
Under-utilized plant capacity
SAIGE - (DATA DRIVEN MODELS)
−47% recovery prediction error
+7–10 pp recovery uplift
+5–8% mill capacity unlocked
Observed in:
Sulfur-constrained systems
Epithermal gold