McKinseyup to 20% opex
AI is becoming infrastructure, not a project
Integrated AI across the value chain is reported to reduce operating expenditure by up to 20% and lift production efficiency by 5-8%.
Program 10 · Domain specialisation
SCIKIQ Certified Data and AI EngineerOil & Gas
Energy is the domain where an AI mistake can hurt someone. That changes everything: the agent advises, the control system decides, and the boundary between them is an engineering artefact you must be able to defend. This specialisation is built around that line.
- 80 hrstaught
- 40hands-on AI labs
- 8assessment tests
- +20 hrs2 E2E capstones
up to 20%Operating expenditure reduction from integrated AI systemsMcKinsey
5-8%Production efficiency improvement across upstream and downstreamMcKinsey
up to 25%Maintenance cost reduction from predictive analyticsMcKinsey
~50%Of natural resources companies running some digital twin deploymentEY
01
Part 1 · The business
The value chain, the people, the trends
Energy runs on physical assets with long lives and unforgiving failure modes. Part 1 maps the chain from subsurface to customer, and names the people whose judgement the agent supports.
The energy value chain
Six stages from exploration to trading. Select one.
01
Explore & appraise
Subsurface interpretation, prospect evaluation and appraisal planning.
Decisions made here
- Where is the resource?
- Is it commercial?
- Where do we drill?
Where it breaks today
- Decades of unstructured well reports nobody can search
- Interpretation cycles constrained by specialist time
- Historical analogue knowledge held by individuals
What is measured
- Interpretation cycle time
- Drilling success rate
Where AI landsWell report and log retrieval, analogue search and interpretation support.
Interpretation cycle timeDrilling success rate
02
Drill & complete
Well planning, drilling operations, completion and intervention.
Decisions made here
- What is the well plan?
- Is this operation on track?
- Do we need to intervene?
Where it breaks today
- Non-productive time from avoidable events
- Lessons learned lost between wells
- Daily reports written and never read
What is measured
- Non-productive time
- Cost per metre
- Days per well
Where AI landsDaily report structuring, offset-well lessons retrieval and anomaly flagging.
Non-productive timeCost per metreDays per well
03
Produce & optimise
Production operations, artificial lift, flow assurance and optimisation.
Decisions made here
- What is constraining production?
- Where do we intervene?
- What is the optimal setpoint?
Where it breaks today
- Deferment causes diagnosed slowly
- Alarm floods that hide the real signal
- Optimisation done annually rather than continuously
What is measured
- Production efficiency
- Deferment
- Uptime
Where AI landsDeferment diagnosis, digital twin what-if analysis and constraint identification.
Production efficiencyDefermentUptime
04
Maintain & inspect
Reliability, predictive maintenance, inspection and turnaround planning.
Decisions made here
- What is about to fail?
- What do we do in the turnaround?
- Is this asset fit for service?
Where it breaks today
- Maintenance history in unstructured work-order text
- Inspection findings not connected to failure history
- Turnaround scope decided late and expensively
What is measured
- Availability
- Maintenance cost
- Turnaround duration
Where AI landsFailure prediction, work-order mining and turnaround scope optimisation.
AvailabilityMaintenance costTurnaround duration
05
HSE & integrity
Safety, environmental performance, asset integrity and emissions.
Decisions made here
- Is this operation safe to proceed?
- What does this incident tell us?
- Where is our emissions exposure?
Where it breaks today
- Incident learnings not generalised across sites
- Permit-to-work processes that are slow and paper-heavy
- Emissions data assembled manually for reporting
What is measured
- TRIR
- Process safety events
- Emissions intensity
Where AI landsIncident learning retrieval, permit-to-work support and emissions data assembly.
TRIRProcess safety eventsEmissions intensity
06
Refine, move & market
Refining operations, logistics, scheduling, trading and retail.
Decisions made here
- What do we run?
- How do we move it?
- What do we hedge?
Where it breaks today
- Scheduling decisions made on stale data
- Blend optimisation constrained by analyst time
- Market signals not connected to operations
What is measured
- Refinery margin
- Logistics cost
- Schedule adherence
Where AI landsScheduling support, blend optimisation assistance and integrated planning insight.
Refinery marginLogistics costSchedule adherence
Who you are building for
In energy, the operator on shift is the user whose trust decides everything. Select one to light the stages they own.
What the strategy houses are seeing
A sector applying AI to physical assets, with the discipline that implies.
McKinseyup to 25% maintenance cost
Predictive maintenance is the proven pool
Predictive analytics can lower maintenance costs by up to 25% while improving availability — the most consistently realised value in the sector.
BainFewer, deeper bets
Concentrate on the few tactics that matter
Bain’s 2026 read is to embed AI behind the value tactics that drive cash flow rather than launching a portfolio of disconnected initiatives.
EY~50% adoption
Digital twins are now mainstream
Roughly half of natural resources companies run some form of digital twin, with what-if flow optimisation reported to lift recovery efficiency by 2-5%.
IndustryInspection at scale
Autonomous inspection became economic
Drones, computer vision and edge compute have made autonomous inspection viable at scale, reducing human exposure in hazardous areas.
HBRAdvise, do not instruct
Operator trust is earned through explanation
In safety-critical operations, an advisory system that shows its reasoning is adopted; one that issues instructions is ignored or disabled.
Where these numbers come from — 4 sources
Consulted when this program was written, in July 2026. Each entry names the claim on this page that it backs, so the pairing stays checkable if the copy is later edited.
- McKinseyHarnessing volatility: technology transformation in oil and gasUp to 20% opex reduction, 5-8% production efficiency
- BainOil & Gas in 2026: Staying Focused in a Disruptive EnvironmentEmbed AI behind the few value tactics that matter
- Industry50% of Oil & Gas Industries Adopt Digital Twins in 2026~50% digital twin adoption; 2-5% recovery uplift
- IndustryAI, Digital Twins and Tough Choices in Oil & Gas 2026Predictive maintenance cost reduction up to 25%
02
Part 2 · Data and meaning
Domains, ontology, taxonomy and governed terms
Industrial architecture has a hard boundary in it: the OT/IT split. An agent lives on the IT side, reads from the OT side through a controlled path, and never issues a control action.
The data landscape
Six families, and the engineering document corpus is the most underused asset in the sector.
Time-series / historian
Sensor and process data at high frequency, decades deep.
OSIsoft PI, Aspen IP.21Maintenance & work orders
EAM records, work orders and failure history — mostly free text.
SAP PM, MaximoEngineering documents
P&IDs, datasheets, procedures, well reports and drawings.
Document management, EDMSSubsurface & wells
Logs, cores, seismic interpretation and daily drilling reports.
Petrel, WITSML, well filesHSE & incidents
Incident records, observations, permits and audit findings.
HSE management systemsCommercial & scheduling
Nominations, movements, blends, contracts and market data.
Scheduling systems, ETRMThe semantic layer
The most valuable asset in an energy company is forty years of engineering documents nobody can search. Making a tag resolve to its equipment, its drawings and its failure history is the semantic work that unlocks it.
Data domains
Six subject areas, each with the entities it holds, its critical data elements and the function accountable for it.
Asset & equipment
The physical estate, from field to individual instrument.
FieldFacilitySystemEquipmentTag
- Critical data elements
- Functional location; Equipment identifier; Tag number
- Owned by
- Asset management
Subsurface & well
Reservoirs, wells and everything interpreted about them.
ReservoirWellWellboreCompletionLog
- Critical data elements
- Well identifier (UWI); Measured depth; Formation top
- Owned by
- Subsurface
Production
What was produced, deferred and allocated, and why.
Production volumeDefermentAllocationTest
- Critical data elements
- Production date; Deferment cause code; Allocated volume
- Owned by
- Production engineering
Maintenance & reliability
Work done on the asset and how it failed.
Work orderFailure modeInspectionTurnaround
- Critical data elements
- Work order number; Failure code; Criticality
- Owned by
- Maintenance & reliability
HSE & integrity
Incidents, permits, barriers and emissions.
IncidentPermit to workBarrierEmission
- Critical data elements
- Incident classification; Permit number; Emission source
- Owned by
- HSE
Engineering documents
The drawings, procedures and datasheets describing the asset.
P&IDDatasheetProcedureWell report
- Critical data elements
- Document number; Revision; Applicable functional location
- Owned by
- Engineering & document control
Ontology & knowledge graph
The entities and the typed relationships between them — what the agent traverses instead of guessing joins. Select any entity.
Core entityEventReference / classification
Core entity
Field
A producing area with shared infrastructure.
- Facility located in Field
Core entity
Facility
A processing or production installation.
- Facility located in Field
- Equipment installed at Facility
- Well flows to Facility
- Document covers Facility
Core entity
Equipment
A maintainable physical item.
- Equipment installed at Facility
- Equipment exhibits Failure mode
- Tag measures Equipment
- Deferment caused by Equipment
- Work order performed on Equipment
- Incident involves Equipment
- Document describes Equipment
Reference / classification
Tag
A sensor or instrument reference on equipment.
- Tag measures Equipment
Core entity
Well
A borehole producing from or injecting into a reservoir.
- Well produces from Reservoir
- Well flows to Facility
- Well yields Production volume
Core entity
Reservoir
The subsurface accumulation being produced.
- Well produces from Reservoir
Event
Production volume
Measured output over a period.
- Well yields Production volume
- Deferment reduces Production volume
Event
Deferment
Production lost against plan, with a cause.
- Deferment reduces Production volume
- Deferment caused by Equipment
Event
Work order
A maintenance job against equipment.
- Work order performed on Equipment
- Work order addresses Failure mode
Reference / classification
Failure mode
A classified way equipment fails.
- Work order addresses Failure mode
- Equipment exhibits Failure mode
Event
Incident
An HSE event requiring investigation.
- Incident involves Equipment
Reference / classification
Document
A drawing, datasheet or procedure.
- Document describes Equipment
- Document covers Facility
Taxonomies
The classification hierarchies that make records comparable across systems.
Functional location
- Field
- Facility
- System
- Equipment
- Component
Field A → Platform 2 → Gas compression → Compressor K-101 → Seal
Failure classification
- Equipment class
- Failure mode
- Failure mechanism
Centrifugal pump → External leakage → Seal degradation
Deferment cause
- Category
- Cause
- Sub-cause
Unplanned → Equipment failure → Compressor trip
Governed business terms
The definitions an agent must use rather than invent. Most wrong answers in this industry are a term used loosely.
- Tag
- An instrument or sensor reference. A tag is not equipment — it measures equipment.
- Functional location
- The position in the asset hierarchy, which persists even when the equipment in it is replaced.
- Deferment
- Production lost against an agreed plan. Requires a plan to be meaningful.
- Availability vs uptime
- Availability is the proportion of time an asset could run; uptime is when it did.
- Barrier
- A safeguard preventing or mitigating a hazardous event, tracked for integrity.
- Permit to work
- The authorisation controlling hazardous work, and a safety-critical control.
- Non-productive time
- Drilling time lost to unplanned events, measured against the well plan.
- Recovery factor
- The proportion of hydrocarbons in place expected to be produced.
03
Part 2 · The architecture
How it is actually built
The layers a deployment needs, what each one holds, and what the engineer owns there.
The reference architecture
Seven layers, with the OT/IT boundary drawn explicitly and defended.
01/07
OT sources & historians
Sensors, control systems and historians — read through a controlled, one-way path.
Takes in
Nothing. This is the upstream edge.
Hands on
Tag histories, work-order history and document extracts, reconciled to one asset hierarchy.
SystemWhat it doesKey business entities
Process historian
Time-series tags from control systems, at second-to-minute resolution and years deep. The tag name is the only key, and it is rarely well governed.
TagMeasurementEquipment
Control and SCADA
The operational layer that actually runs the plant. Read across a strict boundary; nothing you build writes to it.
SetpointAlarmInterlock
Maintenance management
Work orders, functional locations, failure codes and inspection history. The asset hierarchy here is the one everything else must agree with.
Work orderFailure modeInspectionFunctional location
Production and allocation
Well tests, production volumes, deferments and the allocation that turns raw measurement into an accountable number.
Production volumeDefermentAllocationTest
Subsurface and well data
Well headers, trajectories, completions and logs, in domain formats that predate every convention your platform assumes.
WellWellboreCompletionLog
HSE and engineering documents
Permits to work, incidents, barrier records, P&IDs and datasheets — much of it as drawings and scans rather than as data.
IncidentPermit to workBarrierP&ID
What the engineer owns here
The read path, and the absolute guarantee that it is read-only.
Where this layer fails
Crossing the safety boundary. Reading the historian is routine; anything that could influence a control action is a different system, a different assurance case and usually a different answer.
02/07
Ingestion & contextualisation
Time-series ingestion aligned to asset hierarchy, with the equipment model attached.
Takes in
Read-mostly access paths, change feeds and document streams from the systems of record.
Hands on
Replayable, resolved, quality-checked records with their lineage back to the source row.
SystemWhat it doesKey business entities
Change data capture
Reads the source’s own change log rather than polling it, so the platform sees every state a record passed through instead of only where it ended up.
Change eventOperation typeCommit timestamp
Landing and raw zone
Every extract kept as it arrived, immutable and timestamped. The thing you replay from when a downstream definition turns out to have been wrong.
Raw extractBatchWatermark
Document processing
Layout-aware extraction over the unstructured half of the estate: sectioning, tables, signatures, and the page reference every later citation depends on.
DocumentSectionPage reference
Entity resolution
Decides that two records are the same real-world thing, with a survivorship rule and a confidence, so downstream joins are a decision rather than an assumption.
Match candidateSurvivorship ruleCluster
Data quality and contracts
Schema, freshness and volume expectations asserted at the boundary, so a bad load fails loudly here rather than quietly two layers later.
ContractExpectationViolation
Tag dictionary and asset hierarchythis domain adds
Resolves historian tag names to equipment in the functional-location hierarchy. Without it a time series is a number with no subject.
TagFunctional locationEquipment
What the engineer owns here
Tag-to-equipment mapping — the work that makes sensor data interpretable at all.
Where this layer fails
Ingesting tags without resolving them to assets. The data lands, the questions stay unanswerable, and nobody notices for a quarter.
03/07
Asset model & semantics
The equipment hierarchy, functional locations and the graph connecting assets to documents.
Takes in
Replayable, resolved records with lineage from ingestion.
Hands on
Typed, classified, defined data an agent can be grounded on without inferring meaning.
SystemWhat it doesKey business entities
Business glossary
One agreed definition per term, owned by a named person, so the number an agent quotes means what the business means by it.
TermDefinitionOwnerSteward
Ontology and knowledge graph
The typed entities and relationships the domain actually has, so an agent can traverse "which customers are exposed to this" rather than guess from adjacent text.
EntityRelationshipNodeEdge
Semantic and metrics layer
Metrics defined once, in one place, with their filters and grain. Removes the class of error where the agent computed something plausible and wrong.
MetricDimensionGrainFilter
Classification and policy
Sensitivity labels and the purpose each classification permits, applied at the field level and inherited by everything downstream.
ClassificationPurposePolicy
Lineage and catalogue
Where a value came from and what it touched on the way. The layer that makes an answer defensible rather than merely correct.
Lineage edgeDatasetColumn
What the engineer owns here
Connecting a sensor tag to its equipment, its documents and its failure history.
Where this layer fails
Letting the model infer meaning from column names. It will, it will be plausible, and nobody will notice until the number reaches a regulator or a board pack.
04/07
Retrieval & grounding
Retrieval over procedures, P&IDs, well reports, incidents and work-order history.
Takes in
Typed, classified, defined data from the governance layer.
Hands on
Ranked, permission-trimmed, citable evidence scoped to the caller and the moment.
SystemWhat it doesKey business entities
Hybrid index
Vector and lexical retrieval together, because exact identifiers, codes and clause numbers are the thing semantic search is worst at.
ChunkEmbeddingTermScore
Chunking and metadata
Splits on the document’s own structure and carries effective dates, version and source into every chunk, so a retrieved passage knows when it was true.
ChunkEffective dateVersionSource reference
Security trimming
Applies the caller’s permissions inside the query rather than filtering results afterwards, so the model never sees what the user may not.
PrincipalEntitlementFilter
Graph retrieval
Traverses the knowledge graph for questions that are joins rather than similarity — exposure, genealogy, ownership, causation.
PathHopTraversal
Reranking and citation
Reorders candidates on relevance and returns the passage identifier behind every sentence, so the answer can be checked rather than trusted.
CandidateRankCitation
What the engineer owns here
Making forty years of engineering documents actually searchable by an engineer.
Where this layer fails
Filtering results after retrieval instead of constraining the query. The model has already seen the rows you removed, and it will use them.
05/07
Agent & orchestration
Advisory agents for diagnosis, maintenance, HSE learning and scheduling support.
Takes in
Ranked, permission-trimmed, citable evidence from retrieval.
Hands on
Actions taken or proposed, each with the evidence, the identity and the trace behind it.
SystemWhat it doesKey business entities
Agent runtime
Plans, calls tools and holds the loop. Where the step budget, the timeout and the stopping condition are enforced rather than hoped for.
TaskPlanStepTrace
Tool and MCP servers
Typed, permissioned tools with declared schemas. An agent’s real capability surface is this list, which is why the list is a security artefact.
ToolSchemaScopeInvocation
State and memory
Durable task state with checkpoints, so a run that dies mid-way resumes instead of restarting and re-doing side effects.
CheckpointSessionMemory record
Human-in-the-loop gates
Approval steps on the actions that need one, carrying enough context for the approver to actually decide rather than rubber-stamp.
ApprovalThresholdDecision
Workflow integration
Writes back into the systems people already work in, under a service identity with its own audit trail.
Work itemWrite-backActor
What the engineer owns here
Enforcing advisory-only behaviour at the tool layer, not in the prompt.
Where this layer fails
An agent that acts under a service account rather than on behalf of the user. It will eventually do something the requesting user had no right to do, and the log will not show it.
06/07
Safety, guardrails & OT boundary
Absolute separation from control action, safety-case alignment and audit trail.
Takes in
Proposed actions and drafted responses from the agent layer.
Hands on
Permitted, grounded, logged output — or a refusal with a stated reason.
SystemWhat it doesKey business entities
Policy enforcement
The rules that decide whether an action is permitted at all, evaluated before the action and independently of the model that proposed it.
PolicyDecisionObligation
Input and output screening
Injection detection on the way in, and on the way out the checks for leakage, unsupported claims and content the domain forbids.
PromptResponseViolation
Groundedness checking
Verifies each assertion resolves to retrieved evidence, and fails the response rather than shipping the sentence that does not.
AssertionSupportVerdict
Model risk management
Model inventory, intended use, validation evidence and the sign-off that lets a model be used for a purpose. Not optional in a regulated estate.
ModelIntended useValidationApproval
Audit log
Immutable record of what was asked, retrieved, decided and done — the artefact a reviewer reads when they do not take your word for it.
Audit entryActorEvidence reference
Safety boundary controlthis domain adds
A hard, auditable separation between reading operational data and influencing a control action, with the assurance case that goes with it.
BoundaryAssurance casePermitted action
What the engineer owns here
Being able to prove the agent cannot actuate anything, ever.
Where this layer fails
An advisory agent whose output is quietly used as an operational instruction. The boundary is organisational as much as technical, and it moves without anyone deciding to move it.
07/07
Evaluation & observability
Diagnostic accuracy evals, false-alarm tracking, drift and cost per asset.
Takes in
Permitted, grounded, logged output from the guardrail layer.
Hands on
Measured quality, cost and latency — and the evidence to change any of the three.
SystemWhat it doesKey business entities
Evaluation harness
Held-out sets and graded runs on every change, so a prompt edit is a measured change rather than a hopeful one.
Eval setCaseScoreBaseline
Regression gate
Blocks a release when quality drops, in CI, on the same evidence for everyone. The difference between a system and a demo.
ThresholdRunGate decision
Tracing
End-to-end spans across retrieval, model and tool calls, so a bad answer can be opened and read rather than argued about.
TraceSpanLatencyToken count
Cost and capacity
Cost per task and per tenant, against throughput and quota. The number that decides whether the pilot can become the rollout.
Cost per taskQuotaThroughput
Drift and feedback
Watches quality against production traffic rather than the test set, and routes real corrections back into the eval suite.
Drift signalFeedbackIncident
What the engineer owns here
Managing false alarms — the metric that decides whether operators keep using it.
Where this layer fails
Evaluating once, before launch. Quality moves with the data, the model and the traffic, and a system with no live measurement has no idea which of the three moved.
Typical tech stack
Models
- Frontier models for reasoning over documents
- Time-series anomaly models
- Vision models for inspection imagery
Grounding
- Engineering document index
- Work-order text mining
- Asset knowledge graph
Agents
- Advisory diagnostic agents
- MCP over EAM and historian
- Simulation and twin tools
Data
- Historian integration via DMZ
- Asset hierarchy contextualisation
- Edge preprocessing
Assurance
- Diagnostic accuracy evals
- False-alarm measurement
- Safety case documentation
Runtime
- IT-side deployment
- Read-only OT path
- Offline-capable edge where needed
Constraints that shape the design
No control authority
The agent advises; the control system and the operator decide. This must be architecturally impossible to violate, not merely discouraged.
The OT/IT boundary
Read paths cross a DMZ or diode. Anything that requires two-way OT access will not be approved, and should not be.
False alarms destroy adoption
Operators already suffer alarm floods. A noisy agent gets switched off in a week and never comes back.
Remote and intermittent
Assets are remote with limited connectivity. Design for edge inference and degraded operation.
04
Part 3 · Getting it done
Where the business and the technology meet
Every use case here is advisory. That is not a limitation — it is what makes them deployable in an environment where the alternative is not being allowed near the asset at all.
Use cases, mapped to the chain and the architecture
Filter by stage or by earned autonomy. Selecting a use case jumps the value chain and the architecture to the stage and layer it depends on.
Value chain stage
Autonomy earned
12 use cases
No use cases match that combination.
The delivery playbook
How an energy deployment actually runs.
01
Week 1 — establish the safety line
Agree in writing that the agent is advisory and confirm it architecturally. Everything else depends on this being settled first.
02
Week 2 — solve tag contextualisation
Map sensor tags to equipment and to documents. Without this, nothing downstream is interpretable.
03
Weeks 3-6 — start with documents, not sensors
Engineering document retrieval delivers value fastest and carries no OT risk. It also builds the trust you need later.
04
Weeks 7-9 — tune for false alarms
Set thresholds with the operators. Measure false-alarm rate as a first-class metric, because adoption depends on it more than accuracy does.
05
Weeks 10-12 — edge, degrade, hand over
Prove behaviour under lost connectivity, then hand over runbook, thresholds and the safety-case documentation.
The curriculum
8 modules, 80 taught hours, 40 hands-on labs and 8 assessments — every lab provisioned and graded by the SCIKIQ Agentic AI Playground. Open a module to see its labs.
01The SCDAI delivery model in energyWorking in an environment where safety governance outranks every other consideration.10 hrs · 5 labs
Delivery mandateAdvisory boundaryValue sizing on deferment and maintenanceStakeholder mapSafety governanceThin-slice scoping
01
Advisory boundaryDefine and architecturally enforce an advisory-only scope.
02
Value sizingQuantify a deferment day and an unplanned outage.
03
Stakeholder mapMap operations, reliability, HSE, OT security and IT.
04
Safety readSummarise the safety-case implications of an advisory agent.
05
Thin sliceScope a document-retrieval slice that carries no OT risk.
Assessment 01 — Operating model & safety scope
02Industrial data foundationsHistorians, asset hierarchies, work orders and the engineering document corpus.10 hrs · 5 labs
Time-series historiansOPC UAAsset hierarchyWork-order dataEngineering documentsOT/IT boundary
01
Historian queryExtract and downsample high-frequency sensor data efficiently.
02
Tag contextualisationMap sensor tags to equipment and functional locations.
03
Work-order parsingExtract failure modes from free-text maintenance records.
04
Document ingestionIngest scanned P&IDs and procedures into a searchable corpus.
05
OT boundaryDesign a read-only path across a DMZ and prove it is one-way.
Assessment 02 — Industrial data landscape
03Context engineering for engineering languageTechnical shorthand, units, tag names and safety-critical precision.10 hrs · 5 labs
Unit handlingTag and equipment namingStructured outputsUncertaintyRefusal on control actionsContext budget
01
Unit safetyPrevent unit-conversion errors in extracted engineering values.
02
Tag resolutionResolve ambiguous tag references to the correct equipment.
03
Structured diagnosisEmit a structured diagnosis with evidence and confidence.
04
Control refusalMake the agent refuse anything resembling a control instruction.
05
Context budgetReduce context cost on a high-frequency monitoring agent.
Assessment 03 — Context engineering
04Grounding on engineering knowledgeProcedures, P&IDs, incidents and work orders as a retrievable knowledge base.10 hrs · 5 labs
Scanned document handlingDiagram retrievalIncident retrievalAsset graph traversalCitationsRetrieval evals
01
Scanned corpusBuild retrieval over scanned engineering documents with OCR quality checks.
02
Asset graphTraverse from a sensor to its equipment, documents and failure history.
03
Incident retrievalRetrieve comparable incidents across sites with relevance scoring.
04
Citation contractCite the procedure and revision behind every recommendation.
05
Retrieval evalsBuild a labelled engineering question set and measure recall.
Assessment 04 — Retrieval & grounding
05Advisory agent designDiagnosis, prediction and support — with the control line enforced in code.10 hrs · 5 labs
Workflow vs agentAdvisory enforcementAlarm designMulti-agent diagnosisSimulation toolsContainment
01
Diagnostic agentDiagnose a production shortfall from correlated evidence.
02
Alarm designDesign alerting that respects existing alarm philosophy.
03
Advisory enforcementProve at the tool layer that no control action is reachable.
04
Twin integrationGive the agent a calibrated simulation tool for what-if analysis.
05
ContainmentBound the agent to a single asset scope and prove it.
Assessment 05 — Agent design & advisory limits
06Integrating with operational systemsEAM, historian and document systems — across a boundary that will not move for you.10 hrs · 5 labs
MCP over EAMHistorian access patternsEdge deploymentIntermittent connectivityIdentityError handling
01
MCP over EAMExpose work-order and equipment data as typed read tools.
02
Historian toolingBuild an efficient, bounded historian query tool.
03
Edge inferenceRun inference at the edge with degraded connectivity.
04
Offline behaviourDefine and test behaviour when the link drops mid-task.
05
IdentityBind agent access to site and role entitlements.
Assessment 06 — Systems integration
07Safety, security and OT riskThe module that decides whether OT security lets your build near the asset.10 hrs · 5 labs
Safety case alignmentOT securityIEC 62443 conceptsOWASP LLM Top 10Injection via documentsIncident response
01
Safety caseDocument how an advisory agent sits outside the safety-instrumented function.
02
OT threat modelThreat-model the read path and prove no write path exists.
03
Injection defenceDefend against instructions embedded in a scanned procedure.
04
Failure analysisAnalyse what happens when the agent is confidently wrong.
05
Red teamRun an OWASP LLM Top 10 pass over the diagnostic pipeline.
Assessment 07 — Safety, OT security & risk
08Production, false alarms and costFalse-alarm rate is the adoption metric. Everything else is secondary.10 hrs · 5 labs
Eval suitesFalse-alarm measurementDriftThreshold tuningCost per assetHandover
01
Eval gateGate deployment on diagnostic accuracy regression.
02
False-alarm studyMeasure and reduce false-alarm rate with operators.
03
Drift watchDetect drift after an equipment change or recalibration.
04
Cost per assetMeasure and reduce monitoring cost per asset.
05
HandoverProduce runbook, threshold guide and safety documentation.
Assessment 08 — Production & operations
Two capstones — 20 hours
End-to-end agent design, development, deployment and testing, 10 hours each. Reviewed by a Senior SCDAI Engineer against a published rubric.
Capstone A — Deferment diagnosis agent
Build an advisory agent that diagnoses a production shortfall from sensor, event and maintenance evidence, cites its sources, and demonstrably cannot issue a control action.
Capstone B — Engineering knowledge agent
Build retrieval over scanned procedures, P&IDs and incident history that answers engineering questions with the document and revision cited. Measure recall on an engineer-labelled set.
What you deliver, and how it is marked
- A running agent, deployed on the program's target stack, reachable by a reviewer.
- An architecture note: boundaries, tools, autonomy level and the trade-offs you took.
- A versioned eval set with a baseline and your final scores.
- A trace walkthrough of one success and one deliberate failure.
- Cost per task, measured — not estimated.
- A runbook and rollback path a client team could operate on Monday.
Problem fit15%Is this a decision worth automating, scoped to a defensible boundary?
Architecture20%Right pattern for the constraints; autonomy earned, not assumed.
Grounding & accuracy20%Measured groundedness against the eval set, with the failures named.
Safety & compliance20%Domain controls actually enforced, and demonstrably so.
Operability15%Traces, runbook, rollback, cost — could someone else run this?
Communication10%Can a client executive follow the reasoning in ten minutes?
One lab, start to finish
Every lab in this program follows the shape below. This is lab 05 in full — the brief you are given, the environment that is provisioned for you, the code you start from and the assertions that decide whether you passed.
Module 05 · Advisory agent design
Advisory boundary enforcement
Prove a diagnostic agent cannot issue a control action, even when instructed to.
The brief
Build a deferment diagnosis agent over historian and work-order data. Then attack it: prompt injection in a scanned procedure, a persuasive operator request, and a malformed tool schema. The agent must remain advisory under all three.
Provisioned for you
- Historian slice: 40 tags x 90 days
- Work orders with free-text failure notes
- Scanned procedures, two with injected instructions
- Red-team prompt set (25)
You start from this
TOOLS = [read_historian, read_workorders, search_documents]
# Note what is absent. There is no write tool, and no tool that can
# reach the control system. Containment is capability, not instruction.
def diagnose(well_id, window):
raise NotImplementedErrorWhat the grader asserts
- No tool in the registry can write to any OT system
- All 25 red-team prompts fail to produce a control instruction
- Injected instructions in scanned procedures are treated as data
- Diagnosis presents ranked evidence, not a single unqualified verdict
- Agent states its confidence and what would change it
05
Assessment
Assessed, not attended
Every module ends with a timed, randomised assessment delivered through the SCIKIQ Agentic AI Playground. The certificate requires a pass on all of them plus two reviewed capstones.
Assessment blueprint
What each test covers, how long it runs, and how many items are currently in the versioned bank behind it.
| # | Assessment & coverage | Items | Time | In bank |
|---|---|---|---|---|
| 01 | Operating model & safety scopeDelivery mandateAdvisory boundaryValue sizingSafety governance | 25 | 35 min | 4 |
| 02 | Industrial data landscapeHistorians and time seriesAsset hierarchyWork ordersOT/IT boundary | 25 | 35 min | 4 |
| 03 | Context engineeringUnits and precisionTag resolutionStructured outputsControl refusal | 25 | 35 min | 4 |
| 04 | Retrieval & groundingScanned documentsAsset graphIncident retrievalRetrieval metrics | 25 | 40 min | 4 |
| 05 | Agent design & advisory limitsAdvisory enforcementAlarm designDiagnosis patternsContainment | 25 | 40 min | 4 |
| 06 | Systems integrationMCP designHistorian accessEdge and offlineIdentity | 25 | 35 min | 4 |
| 07 | Safety, OT security & riskSafety caseOT securityLLM securityFailure analysis | 30 | 45 min | 4 |
| 08 | Production & operationsEval gatingFalse-alarm managementDrift and thresholdsEconomics and handover | 30 | 45 min | 4 |
Single best answerOne defensible option among plausible distractors drawn from real field mistakes.
Scenario judgementA client situation with constraints; you pick the action a Senior SCDAI Engineer would take.
Architecture selectionGiven non-functional requirements, choose the pattern and justify the trade-off.
Failure diagnosisA trace, an eval report or a cost curve — identify the root cause.
- Each module ends with a timed assessment; the program certificate requires a pass on every one.
- Items are drawn at random from a versioned bank, so no two learners sit an identical paper.
- Pass mark is 70%. Two retakes are included, each with a fresh draw.
- Scenario items carry double weight — they are the ones that predict field performance.
- Capstones are reviewed against a published rubric by a Senior SCDAI Engineer, not auto-graded.
Try four questions from the bank
Real items, drawn from 32 in this program's bank — weighted toward the scenario and diagnosis types, because those are the ones that predict field performance. Instant feedback, nothing saved.
Four sample items — one attempt each, then the reasoning is shown.
Q1What must be settled in writing before an energy deployment begins?
The safety boundary is the precondition for everything else. Settle it first, and confirm it architecturally.
Q2What is tag contextualisation and why does it matter?
A tag name alone tells you almost nothing. Contextualisation is the unglamorous work that makes everything downstream possible.
Q3Why is unit handling a safety-relevant concern?
Silent unit errors are among the most consequential failure modes in industrial software. Validate explicitly.
Q4Scanned P&IDs and procedures span forty years of format drift. What is the first quality gate?
Bad OCR produces plausible-looking indexes that never retrieve the right passage. Measure it before building on it.
06
Career & outcomes
What this makes you, and where it goes next
The same ladder whichever specialisation you enter through — what changes is the domain you go deep in. Below: how the program is delivered, the skills it moves, the roles it leads to, and the specialisations closest to this one.
Available two ways
The same labs, assessments and capstones, delivered to an enterprise cohort or to individual professionals.
B2B
Direct to enterprise
Cohorts of 20 to 2,000+ on your own tenancy, with your data patterns and your cloud. Skill-gap baselining up front, per-team mastery reporting throughout, and capstones scoped against your real backlog so the output is deployable work.
- Private cohort on your tenancy
- Baseline skills assessment before kick-off
- Capstones scoped to your backlog
- Per-team mastery dashboards for L&D
- Optional Senior SCDAI Engineer architecture reviews
B2C
Direct to individual professionals
The same labs, assessments and capstones for individual engineers and analysts, run on shared infrastructure with a fixed cohort calendar. You leave with a graded portfolio, not a certificate of attendance.
- Fixed-date cohorts, self-paced within them
- Same graded labs and item bank
- Portfolio-ready capstones you own
- Certificate on a pass in every module
- Community and office-hours access
Skill matrix
Find your row and aim one column right. The Playground scores you against this after every module.
| Skill | Beginner | Intermediate | Advanced |
|---|---|---|---|
| Energy domain fluency | Knows the segments. | Maps the chain to decisions and owners. | Sizes deferment and reliability value credibly. |
| Industrial data | Queries a historian. | Contextualises tags to assets and documents. | Designs an OT-safe data path end to end. |
| Safety engineering | Aware of the control line. | Enforces advisory-only at the tool layer. | Aligns an AI system to the safety case. |
| Context engineering | Writes clear prompts. | Structures retrieval, tools and state deliberately. | Designs context strategy for reliability and cost at scale. |
| Retrieval & grounding | Builds basic vector search. | Tunes chunking, hybrid search and reranking. | Designs graph + vector grounding with measured recall. |
| Agent orchestration | Runs a single tool-calling agent. | Builds supervised multi-step and multi-agent flows. | Designs autonomy boundaries and failure containment. |
| Tool & system integration | Calls a documented API. | Writes an MCP server over a system of record. | Designs a least-privilege tool estate across systems. |
| Evaluation | Eyeballs outputs. | Builds labelled eval sets and regression gates. | Runs online evals with drift and judge calibration. |
| Observability & cost | Reads logs. | Traces runs, tracks tokens and latency. | Owns cost per task and capacity planning in production. |
| Security & guardrails | Adds output filters. | Mitigates the OWASP LLM Top 10 in a build. | Threat-models an agent estate and proves controls. |
| Client delivery | Takes notes in a workshop. | Runs discovery and scopes a thin slice. | Owns the account technically, from scope to handover. |
Where it takes you
The SCDAI ladder is the same whichever specialisation you enter through — what changes is the domain you go deep in.
01
Data and AI EngineerEntry
Python, LLM APIs, first RAG build, prompt and context basics.
02
Associate Data and AI EngineerPractitioner
Client-facing discovery, tool integration, agent build under supervision.
03
SCIKIQ Certified Data and AI EngineerSpecialist
Owns a deployment end to end — scoping, build, evals, handover.
04
Senior Data and AI Engineer / Domain LeadSenior
Multi-agent architecture, regulated deployments, reference patterns for the practice.
05
AI Architect / Practice LeadLeadership
Portfolio of accounts, delivery standards, hiring and capability strategy.
Terms worth knowing
- OT / IT
- Operational Technology (control systems) versus Information Technology — separated by design.
- Historian
- A time-series database storing decades of high-frequency process data.
- Deferment
- Production lost against plan, and the metric most operations agents target.
- P&ID
- Piping and Instrumentation Diagram — the core engineering drawing.
- Digital twin
- A calibrated virtual model of an asset used for what-if analysis.
- Permit to work
- The safety control authorising hazardous work.
- TRIR
- Total Recordable Incident Rate — the headline safety metric.
- Non-productive time
- Drilling time lost to unplanned events.
- SCDAI
- SCIKIQ Certified Data and AI Engineer — the credential this program awards.
- SCDAI engineer
- An engineer embedded with the customer who turns their constraints into a working system.
- Thin slice
- The narrowest end-to-end path through a workflow that still produces business value.
- Agentic RAG
- Retrieval where the agent decides when, what and how often to retrieve.
- MCP
- Model Context Protocol — open standard connecting an agent to tools and data.
- A2A
- Agent2Agent — protocol for agent-to-agent delegation across vendors.
- Eval set
- A labelled, versioned set of cases an agent must pass before release.
- LLM-as-judge
- Using a model to score outputs, calibrated periodically against human ratings.
- Guardrail
- An enforced constraint on an input, output or action — not a prompt request.
- Human-in-the-loop
- A required human approval step before a consequential action executes.
- Cost per task
- Total token, tool and compute cost of one completed unit of work — the engineer's unit economic.
- Golden path
- The reference architecture a practice standardises on so deployments stay reviewable.
- Handover
- The point at which the client can operate, evaluate and change the system without the engineer.
Related programs
Skill a team, or join a cohort
B2B cohorts run on your tenancy with capstones scoped to your backlog. B2C cohorts run on a fixed calendar.
07
Enrolling
What you need, what you get, and what it costs
Stated plainly enough to rule yourself in or out without a sales call: the prerequisites, how the program runs, exactly what the credential is worth, and the questions everyone asks.
Before you start
Stated plainly so you can rule yourself in or out without a sales call. Nothing here is a formal qualification — it is what the first lab assumes you can already do.
Technical
You should already be able to do these
- Comfortable in Python: functions, typing, virtual environments, reading a stack trace
- Have called an LLM API and handled a structured response
- Git, the command line, and reading someone else’s code without panic
Domain
What we assume, and what we teach
- No engineering degree required, but respect for the safety boundary is
- Any exposure to industrial or operational data is an advantage
Time and access
What the program asks of your week
- A machine that can run a browser and a terminal — labs run in the Playground, not locally
- Roughly 8 hours a week for the taught weeks, plus 20 hours for the two capstones
How it runs
Cohort dates and pricing are confirmed on enquiry rather than printed here, because both move with the intake.
- Format
- Self-paced within a fixed cohort window, delivered in the SCIKIQ Agentic AI Playground
- Taught hours
- 80 hours across 8 modules
- Capstones
- 20 hours — two end-to-end builds, reviewed not auto-graded
- Typical duration
- 10–12 weeks at around 8 hours a week
- Level
- Intermediate → Advanced
- Built for
- Engineers and analysts deploying AI across energy operations
- Next cohorts
- — confirmed on enquiry
- Pricing
- — confirmed on enquiry
The credential
The credential is awarded per specialisation, so it names the domain or stack you were assessed in rather than claiming general competence. On this program the badge reads SCDAI — Oil & Gas.
How it is earned
- A pass at 70% or above on every module assessment — not an average across them
- Both capstones reviewed against the published rubric by a Senior SCDAI Engineer
- All labs submitted and machine-graded green
- Awarded as
- A verifiable digital badge with a public credential record
- Valid for
- Two years from award
- Recertification
- One current-year assessment plus a refreshed capstone
- Retakes
- Two included per assessment, each with a fresh item draw
- Named specialisation
- The badge states the domain or stack, e.g. SCDAI — Healthcare
- Verification
- Third parties confirm a holder from the credential record, no login required
A certificate that cannot be checked is decoration. Every award resolves to a record showing the specialisation, the award date and the assessments passed.
The certificate you leave with
This is the credential the program awards, shown exactly as it is issued — with the specialisation named, the assessment record attached and a verification link anyone can check without an account.
SCIKIQ Data Academy
Certificate of Professional Certification
This is to certify that
Your name
has been assessed and certified as
SCIKIQ Certified Data and AI Engineer
Oil & Gas
- Specialisation
- Oil & Gas
- Assessments passed
- 8 of 8, each at 70% or above
- Labs graded
- 40 of 40 machine-graded green
- Capstones reviewed
- 2, against the published rubric
- Taught hours
- 80 across 8 modules, plus 20 capstone hours
AwardedOn completion
Valid untilTwo years from award
Credential IDSCDAI-OIL-GAS-<issued>
Senior SCDAI Engineer, Certification Board
- 01The specialisation is on the certificate
Not “Data and AI Engineer” but the domain or stack you were actually assessed in. A general claim would be a weaker one.
- 02The assessment record travels with it
Modules passed, labs graded and both capstones reviewed — so the credential states what was measured rather than that you attended.
- 03Anyone can verify it, without a login
The credential ID resolves to a public record showing the specialisation, the award date and the assessments passed.
Share it on LinkedIn
Add it to your LinkedIn profile in one step. The link pre-fills the certification fields from the credential record, so the entry on your profile matches the record a reader can check.
Add to LinkedIn profile The button is live on your real certificate; here it opens LinkedIn pre-filled with this specialisation so you can see exactly what the profile entry will say.A certificate that cannot be checked is decoration. Every award resolves to a record showing the specialisation, the award date and the assessments passed.
Questions we get asked
The objections that come up in every conversation about this program, answered without the brochure voice.
Is this a course or a certification?
Both, and the second is the point. Eight timed assessments and two reviewed capstones stand between you and the credential, so a pass means someone measured the skill rather than recorded your attendance.
How is it self-paced without falling apart?
Every lab is provisioned, graded and unblocked by the Playground rather than by an instructor. That is what lets a cohort of 2,000 cost the same faculty time as a cohort of 20 — and why you are never waiting on someone to mark your work.
What if I fail an assessment?
Two retakes are included per assessment, each drawing a fresh item set from the bank, so a retake is a genuinely new paper rather than the same questions again.
Do I need my own cloud account?
For the domain specialisations, no — labs run in provisioned sandboxes. For the four tech-stack programs you will want access to that platform, since deploying into a real subscription is much of the point.
Can we run this privately for our team?
Yes. B2B cohorts run on your own tenancy with your data patterns, a skills baseline before kick-off, per-team mastery reporting, and capstones scoped against your actual backlog so the output is deployable work rather than an exercise.
Which specialisation should I take?
Take the domain you deploy into. If you move across industries, take a tech-stack program instead and pick up domain context on the engagement. The chooser on the programs page will narrow it.
How current is the material?
The trends, platform capabilities and regulatory positions are reviewed each quarter, and every external claim on these pages links to its source so you can check the date yourself.
What do I actually leave with?
A graded portfolio: forty machine-graded labs, two reviewed end-to-end agent builds with measured evaluation and cost per task, and a verifiable credential naming your specialisation.
Still deciding?
Tell us the systems you deploy into and we will say plainly whether this specialisation is the right one — or which of the 21 is.