Key Takeaways

  • Mining block model building still runs on established commercial platforms: Datamine Studio RM, Maptek Vulcan, Hexagon HxGN MinePlan and Deswik.
  • AI is arriving as an add-on layer, not a replacement. Datamine’s MineScape 2026 release added an AI-enabled assistant and advanced scheduling on top of its existing modelling tools.
  • This is marked “Emerging” deliberately. The core block-modelling engines are mature and largely non-AI; the AI value sits in newer scheduling and optimization modules added to them.
  • When choosing a platform, the question is not “which one has AI” (most are adding it). It is “how mature is the specific AI module you need”.

TL;DR

Build your mining block model in an established platform such as Datamine Studio RM, Maptek Vulcan, Hexagon HxGN MinePlan or Deswik. Then look one layer up, at the AI-enabled scheduling and optimization modules these vendors now add. That is where the real intelligence gains are today.

What Is a Mining Block Model?

A mining block model is a 3D grid of blocks that represents the deposit. Each block carries attributes such as grade, density, rock type, geotechnical properties and geometallurgical values. Planners use it to decide which blocks are ore and which are waste, and in what order to mine them.

The block model is not itself an AI artifact. It is built much as it has been for decades. Estimation methods (kriging, indicator methods, or increasingly ML regression) fill each block from drill data. The inputs come from your resource model, geotechnical model and geomet model. What has changed recently is what happens after the block model exists.

How Do I Optimize Mining Block Models With AI?

The practical path has three steps: build the model on a mature platform, add AI where it helps, then pick the right layer to invest in.

Building the mining block model on a mature platform

Datamine, Maptek Vulcan, Hexagon HxGN MinePlan and Deswik all compete on the same core capability. They turn raw drilling and sampling data into a usable 3D block model. For the model itself, all four options are mature and largely equivalent. Therefore, pick based on your existing toolchain, licensing and team familiarity.

Adding machine learning to block attributes

ML can help populate the model, not just schedule from it. For example, you can predict geotechnical and recovery variables with machine learning where assays alone do not cover them. Some sites also keep the model current as new data arrives, using dynamic block models. These approaches still feed the same block model engine.

Using AI modules on top of the mining block model

Datamine’s MineScape 2026 release, built for coal and stratigraphic deposits, is a useful example. It kept the underlying modelling engine but added an AI-enabled assistant and a Tactical Scheduler with an optimization solver. The other vendors increasingly differentiate on similar optimization and scheduling layers. If you are chasing AI value, look at these scheduling, optimization and reconciliation modules. That is where the genuine machine-learning work is currently concentrated.

Where the Block Model Fits in Mine Planning

The block model is the shared input for almost every planning decision. Long-term mine planning consumes it to set pit limits and schedules. Short-term plans, grade control and production reporting draw on it too.

As a result, errors in the model spread into every plan built on it. Expect AI scheduling tools to amplify good data and bad data alike. Finally, reconcile the model against actual mill and mine results regularly, so you know how far to trust it.

FAQs

What is a mining block model used for?

A mining block model represents the deposit as 3D blocks with grade and other attributes. Planners use it to classify ore and waste, design pits and build schedules. It is the main input to mine planning software.

Is AI used to build a mining block model?

Partly. The core engines still use established estimation methods such as kriging. However, ML regression is increasingly used to estimate some attributes, and AI modules are being added for scheduling on top of the model.

Which software is best for a mining block model?

Datamine Studio RM, Maptek Vulcan, Hexagon HxGN MinePlan and Deswik are all mature options. Their core block-modelling capability is broadly similar. Choose based on your existing workflow, licensing and team skills.

Why is AI in block modelling marked Emerging?

The block-modelling engines themselves are mature and mostly non-AI. The AI value is concentrated in newer scheduling and optimization modules. Those modules are still maturing, so evaluate each one on its own evidence.

Try It With Geocluster

Which vendor’s AI-scheduling layer is mature enough to trust, and which is marketing copy? Answering that takes real comparative research. The Geocluster Research Harness is built to help teams run that kind of systematic AI-tooling evaluation for mine planning workflows. You can also read how the Geocluster Research Harness works.