Reliability intelligence for process industries

Find the real root causes of failures. Not just the symptoms.

Subatix connects plant data across systems, logs, and engineering knowledge to identify why failures happen and what to fix next.

Built for mills, mines, and refineries running on legacy systems, Excel, and tribal knowledge.

See in actionTalk to the team
↓50%

unplanned downtime, by attacking the recurring failures driving most of it

↓80%

RCA time — from weeks of manual stitching to hours of evidence-linked review

↑1–2 wk

first ranked findings on your data — no integration, no IT project

01The problem

The data exists. The visibility doesn’t. Equipment downtime compounds.

  • HistorianProcess behavior — pressures, flows, temperatures, vibration
  • CMMSFailure history, work orders, maintenance records
  • MESProduction data, downtime classifications, batch records
DATA ZONEDISCONNECTEDFAILURE?WHYANSWER SITS BETWEEN SYSTEMS
  • ReportsRCA documents, inspection findings, daily logs
  • EmailDecisions, escalations, context the system never captures
  • PeopleOperators, engineers, the why behind the data
When systems don’t connect
  • RCA takes weeks
  • Root causes stay unclear
  • Failures repeat
  • Downtime compounds
02The solution

A system that finds what others can’t see.

HISTORIANCMMSMESREPORTS & LOGSEMAIL THREADSENGINEER NOTESRANK 01Bad actor findingCRITICALRANK 02Bad actor findingHIGHRANK 03Bad actor findingHIGHRANK 04Bad actor findingMEDRANK 05Bad actor findingMEDSubatix Intelligence01Connects systems02Contextualizes data03Runs analysis04Finds root causes05Reasons and suggests fixesSEE OUTPUTS →FRAGMENTED INPUTRANKED OUTPUT

Subatix connects operational systems, engineering data, unstructured inputs, and human knowledge — then turns it into clear, prioritised failure root causes with specific corrective actions.

03The output

Not dashboards. Clear priorities.

Subatix tells teams exactly what happened, why, and what to do next. Every finding is traceable and backed by evidence — maintenance records, control system trends, lab data, shift logs, engineering reports.

Subatix Intelligence · ReliabilityEngine
Updated overnight · Ranked by severity
Copper concentrator · 5 active findings on this siteIndicative · sample output
Rank01

Copper concentrator · Flotation Row 4

System · Cell froth depth control / reagent dosing
Reagent dosing variability — recovery loss
Critical
Recovery
−2.1%
Lost Cu
$4.7M
Conf.
88%
Days
42
Evidence
Lab

LIMS — concentrate grade trending below spec for 6 weeks; assay variability tracked but not flagged to ops

Ops

DCS — collector dosing rate manually overridden across 3 shifts; never raised to MOC

Shift

Operator notes mention “froth looking off” repeatedly — never opened as work order

Action →

lock dosing setpoints under MOC · standardise froth assessment SOP · automated grade alarms with escalation rules

Rank02

Copper concentrator · SAG mill SAG-01

System · Discharge grate & liner
Ore hardness mismatch — uncoordinated feed
In RCA
Failures
5×
Downtime
384h
Loss
$3.4M
Conf.
92%
Evidence
Maint

CMMS — grate replacements cluster around hard-ore campaigns; PM strategy never updated for feed variability

Ops

Historian — mill power excursions exceed envelope on the same days; throughput pushed past safe operating window

Eng

Geology shift reports flag hard zones in advance — never linked to mill control or PM scheduling

Action →

coordinate ore feed with geology · widen grate aperture · add power-excursion trigger for pre-emptive inspection

Rank03

Copper concentrator · Ball mill BM-03

System · Trunnion bearing
Lubrication starvation — cold-start sequences
Open
Vibration
1.8× RMS
Downtime
168h
Loss
$890K
Conf.
78%
Evidence
Proc

Vibration trend — recurring spikes correlated with cold-start sequences after weekend shutdowns

Maint

CMMS — three bearing inspections in 14 months; lube oil sample records inconsistent

Shift

Start-up procedure executed inconsistently across shifts — pre-lube cycle skipped under time pressure

Action →

mandatory pre-lube interlock · lube sample protocol enforcement · revise start-up SOP

Rank04

Copper concentrator · Cyclone bank C-201

System · Apex liner / spigot
Slurry density excursions × silica content
Open
Wear rate
2.3×
Defer
24h/mo
Loss
$520K
Conf.
84%
Evidence
Ops

Density logger — running 8% above target during specific ore campaigns to maintain throughput

Lab

Metallurgy — silica content variation tracked, never linked to wear pattern across cyclones

Maint

CMMS — replacement clustering by ore type; spares unprepared for next high-silica campaign

Action →

density control retune · ore type feeds spares planning · campaign-based wear prediction model

Rank05

Copper concentrator · Conveyor CV-450

System · Belt tracking & chute geometry
Chute geometry × moisture variability — resolved
Closed
Spillage
0/mo
Down since
Aug
Saved
$310K/yr
Conf.
94%
Evidence
Shift

Daily reports flagged spillage during wet-ore periods; resolved after chute redesign in Q1

Maint

CMMS — belt tracking adjustments dropped from 12/mo to 1/mo post-fix

Eng

Engineering review confirmed root cause; chute liner replaced with curved profile

Action →

solution validated · monitoring sustained · case learnings filed for sister conveyors

04Deployment

Value in weeks, not months.

We don’t need a data lake, a warehouse, or 12 months of integration work. Raw exports from your existing systems are enough to start. First actionable insights in 2 weeks. Integrated system in a few months.

  • Works on raw exports — CMMS, historian, downtime records. Whatever you already produce.

  • No heavy integration — we land before any IT project starts.

  • First insights in 1–2 weeks — ranked findings on your data, not a generic demo.

  • Then it becomes a continuous layer — updates as new data arrives, your team operates it.

TYPICAL ENTERPRISE INTEGRATION· slow, expensive, uncertainDATA LAKEWAREHOUSEINTEGRATION?MAYBE INSIGHTSDAY 0MONTH 3MONTH 612+ MONTHSSUBATIX· ranked findings from week 1CONTINUOUS VALUE LAYERRAWRANKEDVALUE FROM WEEK 1–2From raw exports to ranked causes — in two weeks.
05Why Subatix works

Designed for real industrial environments.

01

Messy data ready

Works on brownfield reality — inconsistent failure coding, fragmented exports, decades of operational drift. The data your existing tools bounce off.

02

Cross-functional intelligence

Connects historian, CMMS, Excel, logs, documents, and human inputs — the structured and unstructured layers in one analysis.

03

A continuous system

Not a one-off diagnostic. Updates as new data arrives. Your team operates it; we don't come back every quarter to rebuild the analysis.

04

Learns over time

Every failure analysed makes the system sharper on your assets, your patterns, your failure modes. The longer it runs, the more valuable it becomes.

06Our team

..built from real operations, by people who’ve seen ops problems firsthand.

Our team has hands-on experience across heavy industries, with deep understanding of the systems in use and how people work across operations, reliability, and maintenance teams. We’ve seen how failures happen and how teams spend weeks trying to figure them out — though the answers were sitting in the data they already have. We know how to connect the dots across systems in a safe and secured manner, so teams can see the full picture.

CopperGoldAluminumRefiningPetrochemicals
Alexander Berks
Alexander Berks
CEO, ex-McKinsey

Operational excellence, industrial assets and EPC.

alex@subatix.comLinkedIn ↗ Alexander Berks
Vlad Rozhkov
Vlad Rozhkov
CTO, ex-SciveFlow

AI products and applied systems, full-stack development.

vlad@subatix.comLinkedIn ↗ Vlad Rozhkov
Our team

Operational excellence across process industries and utilities.

07Expansion

Starts with reliability. Expands across operations.

Subatix initially focuses on helping reliability teams find root causes of unplanned downtime and production loss. Once deployed, the same system extends across all operations — because all of these problems originate from the same fragmented data foundation.

TODAY · ENTRY

Reliability & root cause

Recurring failures, RCA compression, action prioritisation — the pain that's most visible and most quantifiable. The wedge.

NEXT · ADJACENT

Production losses & maintenance optimisation

Same data, same plant, same trust chain. Throughput loss, backlog risk, PM effectiveness — the next layer of pain, opened by the same engine.

LATER · FULL STACK

Quality, energy, cost

The full operational stack — not from a roadmap, but because every domain originates from the same data fragmentation problem we already solved.

Which failure on your reliability list has been there for two years? And what would it be worth to actually close it?

See in actionTalk to the team
What happens next

A 30-minute call. Two questions. No deck.

See in action