See It Work
Documentation Support Trust Center Partners
See It Work
SYSTEM: OPERATIONAL OT/IT CONNECTORS: 150+ AUTONOMOUS OPERATION: 15+ DAYS GOVERNED AUTONOMY: ENFORCED AUDIT TRAIL: IMMUTABLE INDUSTRIES: ASSET-INTENSIVE & MISSION-CRITICAL DEPLOYMENT: 3-6 MONTHS VIA APEX CONTROL LOOPS: 3,400+ SYSTEM: OPERATIONAL OT/IT CONNECTORS: 150+ AUTONOMOUS OPERATION: 15+ DAYS GOVERNED AUTONOMY: ENFORCED AUDIT TRAIL: IMMUTABLE INDUSTRIES: ASSET-INTENSIVE & MISSION-CRITICAL DEPLOYMENT: 3-6 MONTHS VIA APEX CONTROL LOOPS: 3,400+

The AI Platform for Governed Autonomous Industrial Operations.

From monitoring to autonomous operations. One platform. No migration.

Control Room OPERATIONS
Smart Wizard · Use-Case Suitability Qualified

Qualify the use case and its value.

  • Use case C3 splitter drift → off-spec risk
  • Data readiness Sufficient
  • Operational value High — protect product quality
  • Autonomy path Advisory → Supervised

Suitability confirmed — qualified to proceed.

From a live problem to governed autonomy. One platform, one run.

Book a working session See how it works
01 / 08 · Qualify Autoplay · hover to pause

Trusted by industrial leaders

XMPro customers XMPro customers

Monitor. Predict. Advise. Coordinate. Act. Evidence at every step.

Stop reacting to your operations.
Start orchestrating them.

One platform runs the whole loop, from qualifying the problem to governed, autonomous action. Here's how a day moves through it.

Phase 01

Prepare for Autonomy

Prepare · XMForge

Your morning starts here. Qualify your hardest operational problems.

Frame the decision, test its value and identify what needs to be ready before you build.

Explore XMForge
XMForge: Wizard library and ValueForge entry point.
Smart WizardXMForge

Wizard Library

Choose a guided conversation.

＋ New Wizard
⌕ Search wizardsAllPublishedDrafts
VFPublished
ValueForge

Build the case for an operational decision.

6 steps
EditNew Session ↗
Choose the right wizard

ValueForge starts with problem and value discovery. Other wizards explore causal AI, digital twins or agent suitability in more depth.

MAPublished
MAGS Use Case Suitability

Explore the fit for multi-agent systems.

4 steps
EditNew Session ↗
CFPublished
CausalForge

Frame a causal AI opportunity.

8 steps
EditNew Session ↗
DTPublished
Digital Twin Business Case

Evaluate a digital twin opportunity.

5 steps
EditNew Session ↗
01 / 05Choose a wizard
Autoplay · hover to pause

Prepare · Operational Context Engine

Give every asset and signal shared operational meaning.

Assets, relationships, and constraints become a governed operational knowledge graph, the model people and agents both reason over.

Explore the Context Engine
Operational Context Engine: Connected operational context for the depropaniser.
Operational ContextOperational Context Engine
Operational Context › Sources › Unit 31

Sources register

Everything feeding the model
H
Plant historianTag names and ranges
Healthy
M
Maintenance systemEquipment and hierarchy
Healthy
P
P&ID setProcess connectivity
Current
CSV
Instrument indexUNIT31-instrument-index.csv
Inspect
Upload and preview Not stored

Inspect the source rows first.

TAGSERVICEUNIT
TI3106TRAY 20 TEMPMissing
TI3108TRAY 26 TEMPMissing
PI3101COL TOP PRESSbarg
The file preserves the same gaps as the source.
Continue to mapping
Preview first. Check the mapping and dry run before committing.
Start with a known source

The instrument index contributes tag identifiers, services and engineering units. Its file identity stays attached so the model can trace information back to its origin.

Inspect the rows before importing

The preview exposes missing units in the source itself. Previewing does not store the file or silently fill its gaps; mapping and a dry run come next.

01 / 06Start with existing sources
Autoplay · hover to pause

Prepare · Procedural Intelligence Engine

Capture how your experts actually operate.

Procedures, SOPs, and hard-won operator know-how are linked and verified, so agents reason in your operational language.

Explore Procedural Intelligence
Procedural Intelligence: Controlled procedure and extracted knowledge side by side.
Procedural IntelligenceXMPie
DOC-3101-OP-001 · Depropaniser Unit

Extract SOP

PDFDepropaniser Unit

Operating Procedures

Source selected
SOP TYPE
Auto-classify from the document
✓Semantic cross-validation
Extract SOP
CONTROLLED OPERATING PROCEDURE
Depropaniser Unit

DOC-3101-OP-001

What the procedure contains
  • Operating modes and prerequisites
  • Safety limits and constraints
  • Step-by-step actions
  • Abnormal responses
Original document retained
Use the controlled procedure

Start with the operating procedure your team already uses. The original PDF remains available for checking the extracted result.

Check meaning across the extraction

Semantic cross-validation checks the emerging structure for consistency. It supports the review draft; it does not replace engineering verification.

01 / 07Load the procedure
Autoplay · hover to pause

Phase 02

Execute Autonomy

Connect · XMPro Data Stream Designer

Build data streams that power better decisions.

Bring real-time and historical data together. Add context, analytics and AI, then deliver the result to apps, agents and workflows.

Explore Data Stream Designer
Data Stream Designer: The connected depropaniser data stream.
Data Stream DesignerDataStreams
DEPROPANISER FEED DISTURBANCE DETECTOR

Bring live and historical data together

Live Depropaniser DataOPC UA · Listener
Historic Depropaniser DataTimescale · Context
Join DataTransformation
REAL-TIMEWhat is happening now?

New readings from the control system.

HISTORICALWhat happened before?

Recent records from the historian.

Listen to the live operation

The OPC UA listener brings new depropaniser readings into the stream. Listeners handle arriving data from operational sources.

Bring the recent history with it

The Timescale context provider retrieves preceding records. Comparing now with what came before gives the next calculation its context.

01 / 06Connect live and historical sources
Autoplay · hover to pause

Operate · XMPro Causal AI

Don’t just predict. Understand why.

Rank likely root causes from live plant data, with engineering-grounded explanations to guide the next action.

Explore Causal AI
Causal AI: Highlighted causal investigation and ranked evidence.
Anomaly Root Cause AnalysisCausal AI
C3 SPLITTER · ANOMALY ROOT CAUSE ANALYSIS

Choose the alert to investigate

Example event
Process viewCausal analyticsEvidence review
Operational context
Selected anomalyInvestigate
PRESSURE DROP
Column differential pressure is high.

Start with what changed. Then ask why.

Differential pressureEvent window
Before the eventNow
Understand the cause behind the warning.Inspect process
Start with one specific event

Choose the pressure-drop alert from the open recommendations. The analysis is tied to this event, rather than a generic list of things that can go wrong.

A trend shows the symptom

The rising trace makes the deviation visible. It does not explain which upstream variable caused it; that is the next question.

01 / 05Select an alert
Autoplay · hover to pause

Operate · XMPro FRS

See the impact. Before you act.

Simulate proposed operating changes to compare likely outcomes, expose trade-offs and check limits—before changing the live process.

Explore Front-Running Simulation
Front-Running Simulation: Operating state and forecast with limits, uncertainty and time horizon.
Front-run thisFront-Running Simulation
FRONT-RUNNING SIMULATION · DEPROPANISER

What if the feed turns heavier?

Example scenario
Front-runCandidatesAdvisory view
4h horizon · 40 model runs
Current operating stateModel starting point
OVERHEAD C4+ IMPURITY1.91 mol%Within 2.5 limit
BOTTOMS C3 LOSS2.75 mol%Within 3.0 limit
Feed rate 15.5 t/hFeed propane 64.8 mol%
Define the next four hoursWhat-if
SCENARIO · HEAVIER FEED
Feed propane64.8 → 55.0 mol%
Heavier feedLighter feed
OPERATING GOALKeep both products in spec.
Preserve throughput.
Overhead C4+ ≤ 2.5 mol%Bottoms C3 ≤ 3.0 mol%
Starting state + scenario + goalFront-run this
Change the scenario, not the plant

This scenario tests a heavier feed over four hours, starting from the current operating state. The scenario control changes the model inputs; it does not write to the live process.

Define success before ranking actions

This example requires both product specifications while preserving throughput. An option that fixes quality by sacrificing production can still fail the goal.

01 / 05Set the scenario
Autoplay · hover to pause

Operate · Decision Intelligence

Put the next best action in front of the operator.

Bring live context, AI advice and recommended actions into one role-specific workspace—with the operator in control.

Explore Decision Intelligence
Decision Intelligence: Pump asset, operating state and condition trends.
Asset Analysis · Pump P78Decision Intelligence
ASSET ANALYSIS · PUMP P78

See the asset and the decision together.

Example workspace
Asset overviewAI AdvisorRecommendations
Operator-led
OPERATING STATERunning
PERFORMANCENeeds review
DECISION STATUSOpen recommendation
Asset contextEquipment + live condition
P78Rotating equipment · Pump
Asset contextNameplate · readings · history
Asset typePumpAreaRotating equipmentHistoryMaintenance available
Condition trendsIllustrative traces
Flow performanceView in detail ↗
Recent operating history
Bearing vibrationView in detail ↗
Recent operating history
!
RECOMMENDATION AVAILABLEPump performance needs investigation.
Review →
Keep the asset in the decision

The published application combines an asset model, nameplate information, current readings and history. The operator does not have to interpret a detached chart without knowing which equipment it describes.

Surface the decision that needs attention

The pump performance alert leads into a recommendation for review. An alert identifies a decision to investigate; it is not a confirmed diagnosis or an instruction to change the equipment.

01 / 06See the operating picture
Autoplay · hover to pause

Autonomous · MAGS + APEX

Let agents act within the bounds you set.

Inside explicit authority, MAGS agent teams execute the routine calls and escalate the exceptions, with every step recorded to a Decision Trace.

Explore MAGS + APEX
MAGS + APEX: Independent agent workspaces, decision traces and the shared activity timeline.
Depropaniser Unit 31 · Agent TeamMAGS + APEX
MAGS CONFIGURATION · UNIT 31

Set the mission. Define the limits.

Example sequence
ConfigurationAgent InstancesHistory
Design
Objective FunctionBusiness objective
APEX · OPERATING DESIGN
Protect recovery.
Control operating cost.

Keep product quality inside the approved operating limits.

ValueRecovery benefit − steam cost
Operating objectiveOptimise the C3/C4 separation
TeamDepropaniser Unit 31
Operating PrinciplesRequired
01
Safety takes precedence

Safety constraints override economic optimisation.

02
Writes require authority

Use approved actions and Guardian pre-approval.

03
Exceptions stop the action

Suspend and escalate when a safety rule is triggered.

Compliance requirements and source provenance stay attached.
APEX makes success explicit

APEX is the design and delivery discipline: business value, operating objectives, constraints and authority become an explicit operating contract. MAGS then implements that design in the team configuration.

Safety is not a weighted preference

The source separates objectives from non-negotiable operating principles. Improving recovery or cost does not permit an agent to bypass a safety rule or its authority.

01 / 07Define the operating contract
Autoplay · hover to pause

Engineer your path to governed autonomy.

Request a demo Explore the platform

Industries

See it in your industry.

View all industries

PRODUCTION PROOF

Real operations. Measurable returns.

Customer result · Oil & gas

$16M Saved every year

Global oil and gas supermajor

Intelligent oil well maintenance.

Deployed in three months, the solution achieved over $8 million in calculated value within the first six months.

Reduction in field service trips
18%
Reduction in maintenance planning
95%

Customer result · Mining

$10M Saved every year

Global potash mining company

Keeping underground conveyors moving.

A predictive maintenance solution for over 50 miles of underground conveyors, composed and deployed in just 30 days.

Reduction in conveyor downtime
30%
Saved every month
9,000t

Customer result · Enterprise scale

35+ Use cases built in-house

Major North American miner

Built in-house. Scaled across six sites.

The in-house engineering team independently composed operational, tactical and strategic solutions to monitor and manage diverse critical assets.

Sites with in-house adoption
6
Assets monitored
1,000+

Independent analyst recognition

Industrial AI.
On the analyst radar.

XMPro in research from Gartner and LNS Research.

59+

Research appearances

to

Explore all analyst recognition

Gartner research inclusion does not constitute endorsement.

For your role

See what changes for your team.

Process operators

Know what changed. Know what to do next.

  • See the operating condition in context.
  • Review recommendations and their evidence.
  • Keep control of the next action.
Explore the operator path

Reliability teams

Turn early warnings into planned interventions.

  • Understand the risk behind an asset alert.
  • Plan interventions around production needs.
  • Coordinate maintenance and operations.
Explore reliability solutions

Operations leaders

Make better operating decisions repeatable.

  • Connect decisions across teams and sites.
  • Set clear authority for people and agents.
  • Review outcomes with the evidence attached.
Plan your first decision path

Digital & engineering teams

Build once. Reuse across operations.

  • Connect the systems you already use.
  • Reuse data, context and decision patterns.
  • Build in governance from the start.
Explore the architecture

Partner teams

Turn your expertise into repeatable solutions.

  • Package domain knowledge into solutions.
  • Reuse integrations and decision patterns.
  • Take customer solutions into production.
Explore partnership

Don’t see your role? Talk to us

Start with one decision. Prove the value.

Bring an operating challenge. See where XMPro fits.

Request a demo Explore the platform