The product
A validated model, running live on plant data
Ranked, explained root cause and recommendations, delivered in real time where the decisions happen.
REAL-TIME INDUSTRIAL CAUSAL AI
The why behind the what
Prediction tells you something will happen. Causation tells you why, and what to do about it — running live, where your plant runs.
Prediction Causation Action
THE PLATEAU
Industrial AI is stuck at prediction
Plants are full of models that say something will happen, and copilots that can chat about it. But operators still have to work out why it’s happening, and what to do.
Prediction without causation cannot be safely automated.
CONTROL ROOM · 07:42
Vibration rising, 72% likely trip in 6h
Throughput drift detected on Unit 2
Efficiency below target band
THE BIG IDEA
Causal AI is the missing link between prediction and autonomy.
You can’t trust an agent to act on a plant it doesn’t causally understand. Causal AI turns predictive alerts and copilots into decisions you can act on — and eventually automate.
THE CAUSAL MATURITY MODEL
The value runs through the pivot
Everyone has been investing in the left of this line. The value — and the path to autonomy — runs through the middle.
- 01 Monitor what is happening
- 02 Predict what will happen
- THE PIVOT 03 Understand why real-time causal inference
- 04 Advise what to do
- 05 Autonomize act safely
WHAT IT IS
Real-time causal inference for industrial operations
What it does
Decide, build, validate, operate
From “is causal the right tool?” through capturing process physics, to validation gates, real-time ranked root cause, and what-if scenarios on the same model.
What it means
Trustworthy, repeatable, in your hands
Causal AI becomes something operations teams can run — not a one-off specialist exercise — with engineering and humans at the centre.
Why you can trust it A disciplined approach is what makes that inference trustworthy. The inference is the product; the method is what makes it safe to act on.
CORRELATION VS CAUSATION
Correlation predicts. Causation decides.
Predictive / correlation-basedWhat will happen Causal AIWhy, and what to do
Basis Statistical association Process physics, hazard knowledge, cause & effect
Failure mode Confident but unexplained; breaks on new conditions Explainable; flags when it’s uncertain or incomplete
Operator gets Another alert to interpret A ranked, explained root cause & recommendation
Path to autonomy Blocked — can’t safely act on correlation Enabled — action grounded in understanding
HOW WE MAKE IT TRUSTWORTHY
A method, not a magic button
The inference delivers the value; the disciplined approach is what earns the right to trust it. Explore how the model is built, validated, and kept honest.
Three sources, one gate Engineering knowledge, operating practice and data evidence converge through an Engineering Validation Gate into a model your own experts stand behind. Decide · build · validate · operate A disciplined method: decide if causal is even the right tool, build the model the guided way, validate with human sign-off, then operate it live. Not a black box Open, established methods, engineering-validated — not a button that auto-generates a graph from data.
WHERE IT FITS
The “why” engine between data and action
Causal AI turns live data and engineering knowledge into a trusted why — the input agents need before they can safely decide and act.
DataStreams the real-time execution layer our moat
Operational Context Engine knowledge & ontology
Causal Suitability Wizards guided entry point
THE WHY ENGINE Real-Time Causal AI ranked, explained root cause
THE DIFFERENTIATOR
The algorithm is open. The advantage is where we run it.
The causal math is open to everyone. Running that inference live, on the plant, at the edge is what competitors do not have.
Everyone else
A causal library in a notebook
One historical extract · one dataset · offline · after the fact
XMPro · live in DataStreams
DCS / controlHistorianAlarms / eventsCMMS / ERP
Why,live Governed
decisionedge or cloud
150+ connectorsreal-timeedge or cloud
The math is a commodity. The real-time execution layer is not.
PROOF · ANONYMIZED REFERENCE
Root cause in minutes, on a live unit
Real-time causal root-cause analysis on a continuous distillation & separation unit at a major petrochemical producer, deployed inside the plant control-system layer.
On a live anomaly, it attributes the deviation to ranked causal factors with a confidence measure, and presents an explained corrective recommendation to the operator.
The causal structure was built from process physics and hazard-study cause and effect: a Validated Causal Model, signed off by engineers and operators, and checked against the actions operators actually took.
The foundation that makes safe autonomy possible.
LIVE ANOMALY · SEPARATION UNIT T+00:04
RANKED CAUSAL FACTORS
Feed composition shift 0.94
Reboiler duty deviation 0.61
Ambient / cooling change 0.28
RECOMMENDATION → OPERATOR Adjust feed-preheat setpoint; explained & awaiting operator confirmation.
Anonymized · illustrative values · full detail under NDA
WHAT IT MEANS FOR YOU
Value for every seat in operations
Control-room operators
Faster responseToo many alerts, not enough answers.
Not just an alarm — the reason and the fix, in minutes. Less overload, more confident action.
Process & control engineers
LeverageKnowledge is tribal, re-solved each time.
Your engineering knowledge, encoded once, working every shift. Consistency and retention of expertise.
Reliability & maintenance
Fewer failuresReacting to failures after the fact.
Find the root cause before the failure, across the fleet. Fewer failures, better planning.
Operations executives
De-riskedAI investment, unclear payoff and risk.
The trustworthy foundation for safe autonomous operations, standardised across sites.
GO DEEPER
Explore the detail
Three deep-dives behind the why engine: how the model earns trust, how every action is governed, and what it looks like running on a live unit.
The disciplined method How the model is built, validated by your engineers, and kept honest. Learn more Governed by design The gates, decision trace, and authority levels behind every action. Learn more See it in action Follow one live decision, from ranked root cause to a governed recommendation. Learn more
THE FOUNDATION FOR SAFE AUTONOMY
Correlation predicts. Causation decides.
Know why, in real time, and act with confidence. The foundation for safe autonomous operations, that you own and scale.
