AI-Native Product Design Humans-in-the-Loop Mining Operations

AI does the modelling. Geologists make the call.

At a large mining operation, blending decisions were made manually, three to five times a day, each taking up to an hour — based on knowledge scattered across spreadsheets and individual experience. The opportunity was an AI-powered blending recommender that could consolidate that knowledge and give geologists a tool they'd actually use, not a black box they'd resist.

Blending recommender interface — AI-generated blend recommendations with geologist approval workflow

Approach

Co-design, not design-for. Weekly sessions with the geology team shaped the roadmap directly — their workflow became the spec. The tool generates the recommendation; the geologist makes the final call and feeds the model. Augmentation, not replacement.

Role

Design Lead & PM across a multidisciplinary team — data science, front-end, design, geology SMEs — five time zones, no face-to-face access to end users.

AI-Native Product Design Humans-in-the-Loop UX Co-Design Cross-Functional Leadership Remote Team Delivery Systems Mapping

Blending recommender interface — detail view of live model grades and stockpile recommendations

98% Accuracy on product grade prediction

$1.5–2M Projected annual impact

70% Adoption in a four-week trial

95% Blend feedback rate

3% Annual blend compliance improvement

"A product planted without roots doesn't grow. The build is the beginning, not the end."

We're all still working this out. If you need a design leader who's curious, adaptive, and wants to help shape what AI design actually looks like — I'd love to connect.

Get in touch →