Agentic AI mineral exploration

From data
to discovery

AI geologists. Mineral systems reasoning. Ranked exploration targets. Every claim cited, every target mapped.

The OreGen workspace during a live run: the agents' activity feed beside the area of interest over 1:100k geology
Pilbara, Western Australia · Lithium exploration · a live run in the OreGen workspace

See OreGen at work

Two real runs, unedited. Watch the team work, then download what it produced.

The Exploration Manager sets questions for the block, four specialists answer with cited claims, and the two targets take shape on the map.

A mature district with 24 operating mines fenced off, the specialists working the open ground between them, and two targets ranked.

How it works

One mandate in, ranked targets out, with the working shown at each stage.

The geological model

Grounded in mineral systems analysis. Every run starts from a mineral systems model of the deposit class, built on four ingredients:

  • Sourceswhere the metals and fluids came from
  • Pathwaysthe routes they travelled
  • Trapswhat held them in place
  • Preservationwhat has kept the deposit since

The specialists test the evidence for each ingredient. Targets are ranked on how strongly the ingredients coincide, and the evidence still missing is stated for each one.

The agent workflow
  1. Stage 1

    Exploration mandate

    You state commodity, area and depth. The manager turns that into a mineral systems model of the deposit class, reads what is already known, and fences off other people’s ground.

    Led by Exploration Manager
  2. Stage 2

    Specialist investigation

    Four specialists test each ingredient of the model, answering the manager’s questions with claims tied to a source and a place on the map.

    Draws on open-file reports and geoscience layers
  3. Stage 3

    Evidence review

    Every claim is weighed for, against or cannot tell, and candidates are critiqued before anything is ranked.

    Produces a claims ledger
  4. Stage 4

    Ranked targets

    Targets with coordinates, the field work that would confirm or kill each one, and a report with every source.

    Produces PDF report and maps
Specialists

Mineral systems geologist · Geochemist · Geophysicist · Structural geologist

Shared data and tools

Exploration reports · books · academic papers · geology, geophysics and geochemistry layers · tenure · prospectivity model · claims ledger

Applications

Three jobs teams give OreGen.

Generate targets on open ground

Find under-explored tracts with the right ingredients, and surface what the legacy record already says about them.

Put your own data in context

Set drill, assay and survey results against the regional record to explain anomalies and sharpen the next programme.

Stress-test a thesis

An independent, cited verdict on a junior’s story before committing capital, with the work that would confirm or kill it.

Founder

Deep domain expertise meets frontier AI capability.

Alexey Zakharov
Alexey Zakharov
Founder, OreGen

Strategy and business development at three major mining producers, now building the AI exploration team.

LinkedIn ↗
  • VP Business Development & VP Innovation, Norilsk Nickel — base metals, PGM, gold, iron ore
  • Chief Strategy and Business Development Officer, EuroChem Group AG — potash, phosphates, iron ore
  • Chief Strategy Officer, Siberian Coal Energy — coal
  • Professional data scientist, top 1% on Kaggle

Talk to us

Write to us about a pilot on your ground, a data partnership, or investment.

✉︎ hello@oregen.ai