Mining AI case studies

The challenge

The operation knew it needed to digitise, but not where to start. Systems had grown site by site and vendor by vendor. Data sat in spreadsheets, fleet systems, the mine plan and the heads of experienced people. Previous assessments had taken months, produced long reports and quickly gone out of date.

Leadership wanted a clear answer to three questions: where are we today, where is the value, and what do we do first?

Our approach

QMAI combined a short onsite assessment with AI-assisted analysis, so the team spent its time on site listening and observing rather than writing.

1. 01Structured site interviews. We interviewed people across the value chain, from the face to the processing plant to finance. Interviews were captured and transcribed with consent.
2. 02AI synthesis. QMAI's agents coded every interview against a mining-specific maturity framework covering data, systems, integration, people and process. Themes, pain points and quotes were linked back to source, so every finding was traceable.
3. 03Systems and data mapping. We mapped each system, its data flows and its owners, revealing duplication, manual handoffs and gaps.
4. 04Value-ranked opportunities. Each opportunity was scored on value, effort, risk and readiness, then sequenced into a phased roadmap.
5. 05Validation with leadership. Draft findings were tested with the site leadership team before the roadmap was finalised.

Why it worked

AI handled the heavy lifting of synthesis, so our mining specialists could focus on judgement. The result was faster, more consistent and fully traceable, giving leadership confidence to invest.

The challenge

The operation had optimisation opportunities everywhere, but no shared picture of how its assets, processes and costs connected. Engineering drawings were out of date. Each department optimised its own area, and nobody could say with confidence which lever would move the bottom line most.

Our approach

QMAI used a site visit as the raw material for a living digital model of the operation.

1. 01Capture. Our team walked the site from ore receival to product dispatch, photographing equipment, nameplates, control screens, process flows and field conditions. Short voice notes and operator conversations were recorded alongside.
2. 02AI interpretation. QMAI's vision and language agents read the photos and notes: identifying equipment and its function, reading nameplate data and screen values, and tagging each image to its place in the process.
3. 03Digital value chain. The tagged evidence was assembled into a digital value chain, linking each process stage to its equipment, inputs, outputs, constraints and costs.
4. 04Value driver tree. We built a value driver tree that traces site earnings down to the operational drivers behind them, such as recovery, availability, throughput, reagent use and energy.
5. 05Optimisation. With the model in place, we tested where changes in each driver had the biggest effect on value, and ranked the opportunities.

Why it worked

Photos and conversations are fast to collect and rich in detail. AI turned that unstructured evidence into structure, and our domain expertise turned the structure into a model that ties every operational decision back to value.

The challenge

Mining runs on contracts: mining services, haulage, maintenance, labour hire, consumables and more. Each contract has its own rates, rise-and-fall clauses, KPIs, rebates and penalties. A single contractor invoice can run to hundreds of line items, and checking each one against the contract is slow, so most are checked only at a summary level.

The result is leakage: rates billed above contract, escalations applied early or twice, rebates and KPI abatements never claimed, and duplicate charges that slip through.

Our approach

QMAI applied its Contract Compliance Engine across the client's key contracts and a sample of invoice history.

1. 01Contract intelligence. AI agents read each contract and extracted the commercial terms that matter: schedules of rates, escalation formulas, KPIs, abatements, rebates, payment terms and obligations.
2. 02Invoice matching. Every invoice line was matched to the contract term that should govern it, along with purchase orders and receipts where available.
3. 03Exception detection. The engine flagged rate variances, incorrect escalations, duplicate or split invoices, out-of-scope charges and unclaimed credits.
4. 04Expert review. Our commercial and mining specialists validated each exception, removing false positives and quantifying the recoverable amount.
5. 05Recovery and prevention. We supported the client to raise claims with suppliers, then set up ongoing monitoring so issues are caught before payment.

Why it worked

AI makes it practical to check every line of every invoice against every clause, not just a sample. Mining expertise ensures the findings are right, defensible and worth pursuing, which turns analysis into cash.

The challenge

The company faced two linked problems. First, industrial amnesia: critical knowledge about past decisions, studies and lessons learned was scattered across reports, emails and the memories of people who had moved on. Executives and directors often could not quickly find why a decision had been made, or what had been tried before.

Second, the board was being asked to approve AI investments and oversee AI risk without a shared understanding of what AI could and could not do in mining.

Our approach

QMAI tackled both problems together, because better access to knowledge is one of AI's most practical uses for leaders.

1. 01Knowledge capture. We brought together board papers, study reports, technical reviews, ASX announcements and project post-mortems into a secure, searchable knowledge base, structured around the company's assets and decisions.
2. 02Ask the Miner for leaders. Executives and directors could ask plain-English questions and get answers grounded in the company's own documents, with a link to every source.
3. 03Decision Room. For major decisions, leaders could review the history, assumptions and risks in one place, alongside how similar decisions had played out.
4. 04AI awareness programme. We ran briefings and hands-on workshops for the executive team and board, covering what AI is, where it creates value in mining, the risks, and the questions directors should ask.
5. 05Governance. We helped the company develop an AI policy and an oversight framework suited to its risk appetite.

Why it worked

Leaders learned about AI by using it on their own information, which made the benefits and the limits real. The company now keeps its hard-won knowledge, and its board can govern AI with confidence rather than caution.