Getting decision-grade answers from AI on mining data: what works, what breaks, and how to verify.

Every capital decision in mining rests on a few derived numbers — a grade, a reserve, an AISC, a recovery. The layer underneath those numbers was never industrialised. Will Coetzer and Paul Cronin examine why generic AI returns confidently wrong answers on mining data, what the token cost curve means for teams who are building on it, and what decision-grade actually requires. Includes a live demo: vanilla AI versus Pulse-connected AI on the same question, same data.

Why attend: You want an honest account of where AI helps and where it breaks on real mining data — not a vendor pitch. You will leave with the 90-second test you can run on any vendor tomorrow, a clear view of the cost curve implications for your workflow architecture, and a practical adoption framework starting from where you are now.