Operational memory
Organize episodes by shift, asset, operating mode, conditions and available evidence.
AOIMODULE 07 · PRODUCT VISION
From operational experience to an explainable recommendation
Target design: connect visual evidence, documented decisions and process variables to support operator review of alternatives.
Request a demo
AOI / PRODUCT VISION CONCEPT
AICE + documented decisions + DCS / historian
CAPABILITIESPRODUCT VISION
Operational experience becomes shared decision context.
AOI AI Process Operator is envisioned as a supervised operational learning layer. Its implementation, available data sources, recommendations and metrics must be defined and validated for each process and project. It does not replace the operator or write directly to the DCS.
TARGET CAPABILITIES
01Contextual decision episodes
02Comparison of equivalent conditions
03Alternatives bounded by constraints
04Recommendations with rationale and uncertainty
05Review and response records
06Evaluation of observed outcomes
BROCHURE · MODULE DETAILPRODUCT VISION
An assisted decision layer under operational governance.
The product vision connects AICE visual evidence, recorded decisions and process data so teams can evaluate alternatives with context.
PRODUCT CONCEPT · NO LIVE PROCESS DATA
Review process context before considering an alternative.
ALTERNATIVE FOR REVIEW
Consider stability constraints before changing feed rate.
AICEDCS / HISTORIANRECORDED EXPERIENCE
The operator reviews the evidence and makes the decision.
01
Relate observed conditions to recorded actions and outcomes.
02
Frame explainable recommendations within validated constraints.
03
Keep the operator responsible for review and decisions.
Documented experience
Relate recorded decisions to observed conditions and subsequent process response.
Explainable alternatives
Show rationale, constraints, missing data and uncertainty for review.
Supervised learning
Evaluate recommendations and outcomes with process specialists involved.
OPERATIONAL MODELPROPOSED DECISION CYCLE
A decision cycle
under human review.
Proposed product direction: combine available evidence and recorded experience to support operator review. Data, constraints and outcomes require project validation.
01 · SIGNALSAICE evidence and events · Recorded actions and decisions · DCS / SCADA and historian variables
→02 · AOIReconstructAlign AICE evidence, events, variables and history into reviewable episodes.
→03 · CONTEXTCompareCompare equivalent operating conditions and identify constraints or missing data.
→04 · ACTIONAssistContextualized decision episodes · Alternatives within defined constraints
OPERATING AUTHORITYHUMAN REVIEW · GOVERNED ASSISTANCE
01
Evidence
Visual events, process variables and documented operating context.
02
Bounded alternatives
Recommendations are limited by constraints defined and validated for each process.
03
Operator review
The operator reviews the evidence and retains authority over the decision.
04
Measured outcome
Any learning is evaluated against the observed result and approved process metrics.
Initial product direction: read-only or shadow mode with recommendations for review. No direct DCS commands. Simulation and greater autonomy remain future work subject to separate validation.
NEXT STEPAOI · PRODUCT VISION
Build intelligence
around your operation.
Talk to RTFicial More information 01 · WHY IT MATTERSOPERATING CONTEXT
From operational experience to an explainable recommendation
in the real world.
Operational decisions can lose their context between shifts. The vision for AI Process Operator is to relate observable conditions, recorded actions and process response before presenting an alternative for review.
01
Reconstruct
Align AICE evidence, events, variables and history into reviewable episodes.
02
Compare
Compare equivalent operating conditions and identify constraints or missing data.
03
Assist
Present an explainable alternative for human review, or abstain when evidence is insufficient.
02 · DATA MODELFROM SIGNAL TO DECISION
An operational and traceable
data chain.
The module transforms visual signals and contextual variables into information ready to operate, measure and audit.
INPUTS / SIGNALS
AICE evidence and events
Recorded actions and decisions
DCS / SCADA and historian variables
Shift, asset, material and operating mode
Available laboratory or production outcomes
→
OUTPUTS / INTELLIGENCE
Contextualized decision episodes
Alternatives within defined constraints
Recommendation with rationale and uncertainty
Operator review and decision record
Comparison with observed response
03 · APPLICATIONSWHERE IT FITS
Designed for
concrete use cases.
Start from an operational need, scale into new scenarios and keep evidence connected to the process.
01Grinding: stability and feed rate
02Flotation: recovery and consumption
03Thermal processes: sequencing and coordination
04Water and energy: efficiency with continuity
INTEGRATIONS
AICEDCS / SCADAHistoriansOperating recordsProcess outcomes
04 · MEASUREMENTOPERATIONAL SIGNALS
Measure to
continuously improve.
These indicators can be combined with OT variables, history and business outcomes to build a shared operational view.
01Evidence and recommendation qualityPROPOSED · TO BE VALIDATED
02Appropriate abstentionsPROPOSED · TO BE VALIDATED
03Stability and recovery by casePROPOSED · TO BE VALIDATED
04Energy, water or reagents by casePROPOSED · TO BE VALIDATED