Built on millions of high-quality samples, we rank the ground for six commodities — so your next drilling program goes where the evidence points. Our AI isn't a black box: it reasons like a geologist, so you can see why every target ranks where it does.
Two ways to do it: a custom model trained on your own data, or mineBAG, which explains every target. We'll help you pick.
Most targeting tools hand you a coloured map and leave you to trust it. We don't work that way.
Most tools give you a bright colour and no explanation. You can't take that to your board or your JV partner.
Satellite alteration maps show what's on the surface — not where the deposit actually is.
None of them tell you the one thing that saves money: where the map is reliable, and where it's just guessing.
Usually within days — in formats your GIS and your geologist can both open.
Every promising spot on your ground, scored and ranked from best to worst.
Host rock, geophysics, geochemistry, alteration — laid out so your team can check our work.
Solid evidence, thin evidence, or a guess — so you know which targets to back and which to skip.
Your best targets first, balancing how promising each one is against how sure we are.
Plain English for your board, your JV partner and your geologist — every line backed by the evidence.
Your data stays yours.
Anything you send us is used for your job only — never put into anyone else's model.
For some ground a custom model scores higher. Where you need to defend every target — or there isn't much local data — mineBAG wins. We'll tell you which fits your ground.
A model tuned to your ground — your choice of method (random forest, gradient boosting, YDF) — trained on satellite imagery, the national geoscience dataset and your own samples.
When: Best when there's enough local data to train on, and accuracy is what matters most.
How Custom ML model worksOur analog engine. Every target comes with a plain reasoning chain and a flag for how much real evidence is behind it.
When: Best when you need to defend every target to a board or JV partner, or there isn't much local data.
How mineBAG worksBoth run on the same national dataset, both are tested the same honest way, and both flag the ground where they're guessing. Not sure which you need? That's the first thing we work out with you.
We tested mineBAG on real, known deposits it had never seen. The score below is how well it picked them out (0.5 is a coin toss, 1.0 is perfect). South Australia is proven for all six commodities.
| Commodity | SA | QLD | WA | VIC | TAS | NT | NSW |
|---|---|---|---|---|---|---|---|
| Gold orogenic / intrusion-related | 0.81 | 0.82 | 0.73 | — | — | — | ✓ |
| Copper porphyry / IOCG | 0.81 | 0.79 | — | — | — | — | ✓ |
| Silver epithermal / vein | 0.78 | 0.74 | 0.76 | — | — | 0.71 | ✓ |
| Lead-Zinc-Silver SEDEX / VMS | 0.75 | 0.78 | — | 0.76 | — | — | ✓ |
| Nickel-Cobalt magmatic / laterite | 0.76 | — | — | 0.85 | 0.87 | — | ✓ |
| Tin-Tungsten granite-related | 0.74 | — | 0.75 | 0.72 | — | — | ✓ |
An empty cell isn't missing coverage — it's ground we haven't proven yet, so we won't quote a number. What works for one rock type and region won't always work for another, and we'd rather tell you than guess. NSW is where we hold the most data (proven across all six) — ask us for the regional numbers. A custom model is tested the same hidden-deposit way, on your ground.
Every target comes with a plain flag — how much real evidence is behind it. That's how you stop spending a drilling budget on a bright spot that's really just a guess.
We've seen ground like this before. Trust the score.
Not much to go on here. Treat the score as a maybe.
We've never seen ground like this — so we tell you straight, instead of inventing a number.
We'd rather say "we don't know here" than hand you a confident number we can't back up.
A lot of vendors quote accuracy numbers measured on the same ground they trained on. Of course they look good — and they tell you almost nothing about your tenement.
We do it the hard way. We hide the known deposits, build the model without them, then check whether it finds them anyway. That's the only test that tells you how it'll do on your ground.
What works depends on the rock type and the region — not just the commodity. We tell you up front whether your ground is one we've proven, and the map flags the rest. Ask any vendor for their hidden-deposit numbers. If they won't share them, you have your answer.
commodities proven
states + ACT
best score so far
(Nickel-Cobalt · TAS)
South Australia
all commodities
We've pulled together and cleaned the public geoscience for the whole continent — so your ground is scored against the full picture, not a thin local slice.
soil & rock samples
mapped rock units · 7 states
mineral fingerprints
known occurrences & mines
map features you can explore
national geophysics layers
All public data — Geoscience Australia, the state geological surveys, the Bureau of Meteorology and Google Earth Engine. No proprietary lock-in. Kept up to date.
One area at a time. Pilot-friendly, no lock-in. We quote your area up front, so you know the cost before we run anything.
Send us your area and the commodity. We'll tell you straight whether it's ground we've proven, and which method fits.
We build and run your model against the full national dataset on the cloud.
Usually within days: ranked targets, the evidence, the trust flags and a drilling shortlist.
Have your own assays or drilling? Send them and we'll fold them in — for your job only.
Drilling in Australia? Let's run a pilot on your ground, so you can judge the output — not the pitch.
Scope a pilot AOI