Tailings and pit-wall instruments are read in isolation. Permit conditions sit across years of documents. Near-miss reports are filed and never read together. The records are already on site — they are not in one place, and they are not interpreted together.
01
Interpreted monitoring streams
Fuse radar, piezometers, survey and satellite already on the facility into one view, with an alert when the combined signal is unusual.
02
Obligations you can search
Turn permits, conditions and regulatory correspondence into answers that carry the source passage they came from.
03
Learning that comes back
Read the near-miss corpus across tasks and shifts, so recurring conditions reach the crew instead of waiting for a serious event.
On the mine site
Productivity has been falling even as digital spend rises. The site's own records are split across systems, teams and paper. In Europe, overlapping environmental reporting sits on top of that. None of that is a missing dashboard. It is a missing shared record.
ProductivityMcKinsey benchmarks mining as about 28% less productive than a decade ago once ore-grade decline is counted. Overall equipment effectiveness sits at 27% underground and 39% in open pits, against 88–92% in oil and gas, steel and refining.
Operations dataEquipment failures are still diagnosed by piecing together radio calls, equipment logs and shift schedules. Month-end reporting still means gathering data from multiple sources into spreadsheets and checking it for accuracy.
ReportingWater, nature, industrial emissions and soil frameworks each ask for their own evidence pack. Euromines calls the result duplication and gold plating, on top of permitting that can take up to 15 years.
AI spend is rising. Siloed data is why that investment is not paying off yet.
EY finds 21% of miners plan to raise AI budgets by more than 20% in the next year, and names fragmented data as the main limit on return. INSUS works from the instruments and records a site already runs. If the gap is a missing sensor, a maintenance job or an existing system, that is what we will say.
Industry figures
Four external figures, each with its source. Nothing on this page is an estimate we made up. Where we cannot cite a number, we describe the situation in words instead.
Productivity
28%
less productive than a decade ago, once ore-grade decline is accounted for. McKinsey also puts the fall at roughly 3.5% a year over that period.
Source: McKinsey, How digital innovation can improve mining productivity
OEE
27%
overall equipment effectiveness underground, 39% in open pits and 69% in crushing and grinding — against 88–92% in oil and gas, steel and refining.
Source: McKinsey, How digital innovation can improve mining productivity
AI spend
21%
of miners surveyed plan to increase AI budgets by more than 20% over the next 12 months. EY finds siloed data and misalignment with business needs are what currently limit the return.
Source: EY, Top 10 business risks and opportunities for mining and metals in 2026
GISTM
67%
of 836 ICMM member tailings facilities are in full conformance with the Global Industry Standard on Tailings Management. 33% remain partial. Conformance is a continuous process, not an endpoint.
Common problems
None of these shows up as a single alarm. They show up in an instrument stream nobody reads alongside the others, in the audit folder still being assembled, and in a report that was filed and never came back.
Survey markers, inclinometers, radar, piezometers and satellite InSAR typically sit with different specialists on different schedules. At Feijão Dam I, piezometers were read manually about monthly until shortly before failure. The Expert Panel found no single stream detected the collapse that killed 270 people — the gap was interpretation, not missing instruments.
See: Tailings Geotechnical Signal IntegrationWater, Natura 2000, nature restoration, industrial emissions and soil directives each have their own process. There is no shared underlying data layer. That load sits on permitting that Euromines says can take up to 15 years.
See: EU Mining Permitting & ComplianceFree-text reports accumulate faster than a safety team can read them across tasks, locations, equipment and shifts. The precursor conditions are in the corpus, but they are usually only found after a serious event forces a retrospective review.
See: Near-Miss & Incident IntelligenceUse cases
Each is a specific starting point — this tailings facility, this permit corpus, this reporting history — and a next conversation if it is worth your time.
Time-align piezometer, inclinometer, radar, survey and InSAR streams into one reviewable record, with joint anomaly flags and GISTM monitoring evidence for geotechnical review.
See the full use case
Make mining permits, conditions and regulatory obligations searchable with cited answers, so compliance teams can find what is required and where it comes from.
See the full use case
Analyze mining near-miss and incident reports to surface recurring hazard conditions and return relevant past events to the crews facing similar situations.
See the full use caseFAQ
A practical look at what INSUS needs, where it fits and what your team can expect.
The work that fits a software and data company is reading records a site already has: tailings and geotechnical instrument streams read together rather than instrument by instrument, the permit, EIA and regulatory corpus made searchable with cited answers, and the near-miss and incident corpus analysed for recurring conditions. Wear inspection and stockpile volumetrics that need new capture are not the starting point.
Usually not. Tailings signal integration, permitting intelligence and incident learning can all be scoped from existing radar, piezometers, survey, document stores and EHS reporting history. If a signal is genuinely missing, we say so in week one rather than build on a gap.
No system should promise that. The Feijão Dam I Expert Panel found that pre-failure deformations were too small and slow for the real-time systems then in place, and that ground-based radar also produced large false readings from vegetation and moisture. The useful work is interpreting several existing streams together, with honest limits, not claiming a failure will always be seen in time.
No. The people who know the pit wall, the mill and the permit pack stay responsible. The aim is to stop them rebuilding the same picture from radio, paper and isolated instrument files, and to surface an unusual combined signal while there is still time to look.
Whatever the site already keeps that describes the problem you want to solve: historian and equipment logs, shift reports, tailings instrument files, survey, InSAR, incident records, and the environmental or GISTM packs you already file. You do not need every source on day one. We first identify what is available, what is reliable and what is actually required.
Yes. The approach is built around the systems you already run rather than a rip-and-replace stack. Integration is defined during the assessment. Safety interlocks and statutory reporting remain yours.
The first step is the current workflow, the data you already hold and whether the gap is software, a missing record or a maintenance job. That assessment is what tells you if a single-site pilot is sensible. The mining work behind this page is still discovery, not a finished architecture or ROI model.
Yes. A focused starting point is usually the practical path: one tailings facility, one underground or open-pit operation, one EU evidence pack, or one mill line. Measure that, then decide whether to widen.
Contact
