15-Session Programme · 25 Minutes Per Day · Free · For the owners and operators of Nigeria’s energy assets
WEEK 1 OF 3 · THE OWNERSHIP MANDATE AND THE FOUNDATIONS YOU CANNOT SKIP · SESSIONS 1–5 · WORK AT YOUR OWN PACE
Introduction and Chapter 0 · OrientationSESSION 1
The Asset Does Not Read the Presentation and the Reader's Toolkit
Why This Matters
Most asset reviews are comfortable. This programme asks the question that tests that comfort: what evidence would make us stop operating this equipment tomorrow? Over fifteen sessions you will build a signed 90-Day Intelligent Asset Plan for your own company, around one unresolved asset decision you choose today. AI can help your people read records, find inconsistencies and prepare that decision. It cannot carry the decision for you. This programme trains you to tell the difference.
25-Minute Module
0:00 – 5:00
The comfortable review, and the finish line
- Production is holding and the next inspection has a green marker. Now ask: what evidence would make us stop operating this equipment tomorrow? If the room is surprised, the review has been describing activity rather than testing the basis for continued operation.
- THE DOCTRINE
Ownership is uncomfortable. An asset owner receives the benefit of production and carries responsibilities that remain when the meeting ends, the contractor leaves or a software subscription expires.
- The central question: how should leadership decide what to inspect, maintain, repair, replace or retire, and what evidence should justify the decision? Week 1 covers ownership and the foundations, Week 2 intelligence in the field and the first lifecycle choices, Week 3 lifecycle, trust, capital and the leadership mandate. The finish line is a signed 90-Day Intelligent Asset Plan, fourteen sections written one per session, with the Asset Decision Record at its heart.
- House rules: write about your own company; bring a real decision; no value claim without a baseline; no decision without a named owner and authority; report a blocker the same day. The pass mark is 70.
- Said once: this is leadership training on decision governance, not legal, investment or engineering advice. It provides no design calculation, inspection procedure, fitness-for-service assessment, operating instruction or permission to continue operating equipment; the law, licence conditions, applicable standards and competent professional judgement apply to your asset. Standards appear only through their published scopes. The frameworks are my synthesis, not statutory tests, engineering standards or evidence of regulatory approval, and the numerical cases are fictional teaching exercises with stated assumptions, not reported company performance; no commercial company is identified or represented by them.
Ask what would make you stop before you ask what to buy.
5:00 – 10:00
Why this belongs in a Nigerian boardroom
- A gathering system, an offshore facility, a gas plant, a refinery and a terminal do not share one maintenance strategy, but they can share a disciplined approach to evidence. NUPRC's August 2026 workforce statement identified both digitalisation, AI and data analytics, and asset integrity, HSE and decommissioning management among priority competencies, and the Commission published 2026 guidelines on upstream asset integrity, life extension and the adaptation of novel technology.
- Keep the boundary. These are upstream documents, and the book does not extend them to every Nigerian installation. The Petroleum Industry Act 2021 establishes separate upstream and midstream/downstream regulatory institutions; applicability depends on the activity, facility, licence and governing instruments. Maintain a facility-specific obligations register and obtain qualified interpretation where boundaries are uncertain.
- Execution changes too: swamp access, offshore mobilisation windows, currency and delivery for imported equipment, work inside an operating terminal. These are questions to investigate at your asset, not assumptions about every Nigerian business. Where a chapter discusses equipment you do not own, or you serve owners as a contractor, supplier or financier, apply it to the nearest equivalent decision or the owner's decision you serve.
Know which obligations apply to your facility before you quote any of them.
10:00 – 14:00
The owner carries the whole decision
- Organisations often divide responsibility sensibly and then lose the decision between the divisions. Inspection, engineering, operations, procurement and finance may each be doing its job while the asset stays in the same condition.
- A recommendation without resources is not an intervention. An approved purchase without a usable operating window is not readiness. A completed repair without acceptance evidence is not restored assurance.
- THE DOCTRINE
Truthful reporting is not achieved by telling people to embrace AI. Leaders have to make it worthwhile to tell the truth: listening, protecting justified escalation and accepting that a well-founded stop decision can be good performance.
Connect the work without confusing the authorities involved.
14:00 – 21:00
The six frameworks, applied to the asset owner
Framework For the asset owner The question it forces 1. Operating-Intelligence Loop Sense, decide, act, learn, ending in a justified disposition Where does this signal stop? 2. Data-Foundation Precondition Decision first, then evidence; readiness is use-specific Adequate for this decision? 3. Autonomy Ladder Retrieve, draft, recommend or execute a permitted step Which actions are authorised, and who withdraws them? 4. Value-versus-Risk Map Consequence of error beside benefit; legal, protection and engineering conditions are gates What if it is wrong? 5. Sequencing Discipline Bounded problem, simpler comparator, validate, rehearse, then expand What comes before the next commitment? 6. Operations-Excellence Lens Effect on work and decisions; activity kept apart from outcome Activity, outcome or counterfactual claim? - The loop and the foundation. A signal needs a time and equipment identity; a decision a basis and an approver; work an authorised scope and acceptance criteria; learning the result. An absent record is not a record of acceptable condition, and a model estimate cannot quietly become an inspection observation.
- THE DOCTRINE
A language model should not become an informal route for changing a protection setting, accepting a pressure boundary or approving continued operation. Authorise individual actions, not a vague level of trust.
- The lens. Reading more reports is activity; a faster checked decision pack may be an outcome; preventing a particular failure is a counterfactual claim. Do not label all three as savings.
- AT THE ASSET IN THE FIELD · THE WORKED CASE
- In the field. In a fictional onshore operator, a degrading pump is repaired. The loop breaks at learn if the repair never returns its condition findings to the equipment history: the next team inherits the same uncertainty.
IN THE BOARDROOM · THE PARALLEL- In the boardroom. The same pump reaches the board inside a completion percentage. The loop breaks at decide if nobody is named to accept its disposition. The same loop, seen twice.
A graph that never changes a decision has not closed the loop.
21:00 – 25:00
Apply the toolkit to your live question, and Section 1
- Write the short record Chapter 0 asks for: the asset and its required function; what is known and what is uncertain; the feasible actions; the technical authority and the spending authority; the deadline imposed by the condition or work window; and the result that would show the decision was useful.
- Then consider AI: retrieval to find evidence, analytics to spot a pattern, or workflow to stop an accepted action disappearing. If one fits, define that narrow purpose and its acceptance test. If no one will fund an already agreed repair, a better prediction may add little.
- THE DOCTRINE
Bring one unresolved asset decision to the next review. Require its evidence, authority and next executable action. Establish exactly where AI could help, then hold it to that purpose.
- Section 1 of your plan is Our company, the one asset decision we will carry, and where AI could help.
The decision you choose today is the one you will defend in Session 15.
Key Terms
Operating-Intelligence Loop
Sense, decide, act, learn, ending in a justified disposition and findings returned to the equipment history.
Data-Foundation Precondition
Choose the decision first, then establish the evidence it needs. Records adequate for one use may be inadequate for another.
Autonomy Ladder
Retrieve, draft, recommend or execute a permitted step, authorised action by action, with a reviewer and a way to withdraw permission.
Facility-specific obligations register
The owner's record of the obligations that apply to a particular facility, since applicability depends on the activity, facility, licence and governing instruments.
Self-Check — Before You Submit
1. A fictional terminal owner is offered an AI tool to predict tank-floor deterioration. The repair was agreed last year but never funded. What does Chapter 0 suggest?
- A. Buy the tool to strengthen the case
- B. Ask the vendor for a funding module
- C. Treat the unfunded repair as the problem; a better prediction may add little
- D. Wait for the model to confirm deterioration
C. Treat the unfunded repair as the problem; a better prediction may add little — The loop is stuck at act, not sense. Waiting for the model to confirm is tempting but only lengthens the gap between knowing and doing.
2. A model estimate fills a missing wall-thickness measurement. What is that estimate?
- A. A study input, not an inspection observation
- B. An observation once the model is validated
- C. Evidence, if confidence is high
- D. A record of acceptable condition
A. A study input, not an inspection observation — An estimate may help a study, but it is not an observation, however confident the model is.
3. Which statement about NUPRC's 2026 guidelines matches the book?
- A. They apply to every Nigerian installation
- B. They replace the Petroleum Industry Act 2021
- C. They are fitness-for-service standards
- D. They are upstream documents; applicability depends on activity, facility, licence and instruments
D. They are upstream documents; applicability depends on activity, facility, licence and instruments — The Act establishes separate upstream and midstream/downstream institutions. Assuming universal application is the tempting error.
Today’s Assignment
Name your company and its assets, choose the one unresolved asset decision you will carry through this programme, and write the short record Chapter 0 asks for.
- Your company. Its name, the assets it owns, operates or serves, and where they sit (upstream onshore, swamp, shallow water or deepwater; gas processing; pipelines; refining; terminals and storage; or another).
- The decision. One unresolved condition, anomaly or lifecycle choice: asset and required function, what is known and uncertain, feasible actions.
- Authority and deadline. The technical authority and spending authority by role, and the deadline imposed by the condition or work window.
- The stop question. What evidence would make your company stop operating that equipment tomorrow?
- Where AI could help. One narrow purpose with its acceptance test, or why a better prediction would add little. No AI in use yet is normal.
- The honest part. What your company currently gets wrong about this decision, and the result that would show the decision was useful.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 1 — Our company, the one asset decision we will carry, and where AI could help.
How this is marked — the rubric
| Names the company, its assets and where they sit | 10 |
| States one bounded decision with function, known and uncertain, and feasible actions | 30 |
| Names technical and spending authority by role, and a deadline set by condition or work window | 20 |
| Answers the stop question honestly and names a real weakness in the company | 20 |
| Defines one narrow AI purpose with an acceptance test, or justifies none | 20 |
| Total · pass mark 70 | 100 |
Submitting session 1
Write your answer in Word or Google Docs. The first line must read
Write your answer in Word or Google Docs. The first line must read
Session: 1, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.Capabilities Covered
Asking the stop questionThe six toolkit frameworksSeparating technical and spending authorityKeeping the upstream boundaryDefining a narrow AI purpose
Chapters 1 and 2 · The Ownership MandateSESSION 2
Ownership Is a Continuing Obligation
Why This Matters
An asset register makes a physical installation sound stable. In operation it changes: its condition, its service, its modifications and the assumptions behind its design. Today you test two things most companies assume. First, that someone owns the next decision on each material condition, rather than a chain of referrals. Second, that your company can actually identify the asset, from the label in the field to the line in the board paper. Neither test starts with technology. Both decide whether any technology you add later will help, or simply hide the gaps more convincingly.
25-Minute Module
0:00 – 5:00
Define the function before the technology
- A current cost centre and operating licence do not, by themselves, show whether a component is fit for its current service. Nor is an old installation automatically unfit because its calendar age is high. Leadership needs an account of condition, function and evidence.
- Ask what the asset must deliver and under what conditions. A pipeline has a transport duty and a containment duty. A compressor may support production while its auxiliaries support safe and reliable operation. A storage tank's usable volume is not its only important attribute: access, containment, protection, drainage and the ability to inspect it also matter.
- ISO 55001:2024 provides a management-system frame for realising value from assets, with decision-making and lifecycle matters among its published areas of emphasis. The practical step is to translate business objectives into functions and decisions an owner can govern.
- Test it on a proposed production increase. Can utilities, containment, maintenance capacity and the evacuation route sustain the extra volume? A model trained on the main process unit may miss the constraint that sits elsewhere. You need a system view, not an isolated performance curve.
Write the duty before you write the specification.
5:00 – 10:00
Separate responsibilities without losing ownership
Role What it holds Asset owner Sets objectives and funds the arrangements Operator Manages the installation within the authorised envelope Technical authority Establishes or accepts the engineering basis within its competence Contractor Executes agreed work; can be accountable for workmanship and records Independent assurance Tests specified claims - In a small business several roles may be held by one person. The responsibilities still need to be distinguished.
- THE DOCTRINE
Do not allow the phrase contractor responsibility to conceal an owner decision. The owner still decides whether the scope addresses the condition, how the work interfaces with operations and what evidence is required before acceptance.
- A certificate should answer an identified question; it is not a general declaration that everything nearby is satisfactory. And the software supplier should not become the only party able to explain your asset: require usable exports, equipment identities, source links, version history and a way to continue when the application is unavailable.
One person can hold three roles; the record should still show three.
10:00 – 15:00
The gap between knowing and doing
- An anomaly can sit for months without a clear decision. Inspection considers its task complete because the report was issued. Operations considers the item with engineering. Engineering considers it awaiting funding. Each statement may be correct while the overall arrangement remains inadequate.
- Create one accountable decision record: the current disposition, the conditions under which it is valid, the outstanding work and the trigger for escalation. If funding is unavailable, say so plainly and return the unresolved operating decision to the appropriate authority. An unfunded action is not a completed risk treatment.
- A useful early AI application sits here: a document assistant can find repeated references to the same anomaly, actions with no owner and records whose stated conditions have expired. Human review remains necessary, because different descriptions may refer to different equipment. The book's executive application: take the most consequential unresolved condition, not the most visible project, and have operations, engineering and finance write one joint decision paper. Judge any AI proposal against it.
- For the board, show a small set of material decisions: required function, evidence supporting continued operation, material uncertainties, feasible interventions, cash needed and the authority accepting the position. Preserve dissent. Do not compress necessary conditions into one average score. And ask whether the work system can absorb what you approve; money without planners, access or acceptance capacity may create a queue rather than improvement.
- THE DOCTRINE
Name the person who owes the next decision to each material asset condition. A chain of referrals is not an ownership arrangement.
Count decisions owed, not tasks closed.
15:00 – 22:00
Establish what you actually own
- Before asking a model to predict an asset's behaviour, establish that the organisation can identify it. A takeover is an evidence transition before it is a software migration: keep documents promised, received, reviewed and accepted as different states. A data room download is not the acceptance event, and a scan with unreadable revision marks is not a controlled current drawing.
- AT THE ASSET IN THE FIELD · THE WORKED CASE
- In the field. The book's fictional handover: the maintenance database lists two identical pumps; field staff say the standby was changed; stores records a motor replacement; the performance curve describes the original configuration. The correct output is an identity conflict requiring verification, not a combined profile: preserve the separate claims and restrict use until it is closed.
IN THE BOARDROOM · THE PARALLEL- In the boardroom. The same conflict belongs on the management page: identities verified, unresolved conflicts and the decisions affected by missing evidence, with an owner for resolution. A percentage of complete fields says less. An assistant that refuses to combine incompatible versions may be more useful than one that fills every field.
- Build relationships that reflect the plant: facility, system, equipment and component, linked to drawing, work order, inspection location and spare, under stable identifiers. ISO 14224:2016 addresses the collection and exchange of reliability and maintenance data in petroleum-related equipment, including data quality. Start with a verified population for a bounded purpose, such as a set of critical compressors, and write its limits.
- Connect physical and contractual boundaries. A shared export route, power supply or berth can decide whether a local improvement produces a business benefit, and physical dependence is not authority to intervene. Otherwise a model may recommend action on capacity that is contractually unavailable.
- THE DOCTRINE
Verify one critical system from the field label to the board paper. Where the chain breaks, correct the ownership of the information before asking a model to reason across it.
A system that admits what it cannot trust is worth more than one that fills every field.
22:00 – 25:00
Write it down, and Section 2
- Before you close the laptop, write the role that owes the next decision on your asset, and the point where its chain from field label to board paper breaks.
- Keep a short exception register that field personnel recognise, and return accepted corrections to the controlled source, not a presentation or chat history.
- Section 2 of your plan is The required function, who owes the next decision, and the asset verified from field label to board paper.
If the chain breaks at the label, no model downstream can mend it.
Key Terms
Required function
What the asset must deliver and under what conditions, including containment, protection and inspectability, not only output.
Accountable decision record
One record per material condition: current disposition, the conditions under which it is valid, outstanding work and the escalation trigger.
Evidence transition
A takeover treated as a transfer of evidence, with documents promised, received, reviewed and accepted tracked as different states.
Identity conflict
Records that disagree about what an item of equipment is. The claims are preserved and the profile's use restricted until verification closes it.
Self-Check — Before You Submit
1. In a fictional handover, an AI assistant merges the database, stores record and performance curve into one tidy pump profile. What is the right response?
- A. Accept it, since a complete profile helps planning
- B. Treat it as an identity conflict and restrict use until verified
- C. Keep the most recent record only
- D. Ask the assistant for higher confidence
B. Treat it as an identity conflict and restrict use until verified — The book's correct output is an identity conflict requiring verification. The complete profile is the trap: filling every field can hide what cannot yet be trusted.
2. Inspection says its report was issued, operations has referred the anomaly to engineering, and engineering says it awaits funding. What does Chapter 1 require?
- A. A new inspection
- B. A risk-log entry for each function
- C. An AI tool to track the referrals
- D. One accountable decision record and a named person who owes the next decision
D. One accountable decision record and a named person who owes the next decision — A chain of referrals is not an ownership arrangement. A new inspection or a separate log adds activity; a document assistant can help find the repeated references, but it does not replace the record or the named owner. If funding is the block, say so and return the decision to the appropriate authority.
3. A contractor presents a certificate for completed work on a valve. What may the owner conclude?
- A. At most, that it answers the question it was issued for
- B. That nearby equipment is satisfactory
- C. That the owner's acceptance is made
- D. That responsibility has passed to the contractor
A. At most, that it answers the question it was issued for — A certificate answers an identified question. Contractor responsibility for workmanship must not conceal the owner's acceptance decision.
Today’s Assignment
For your Section 1 decision, state the required function, separate the responsibilities, draft its accountable decision record and verify it from field label to board paper.
- Function and system view. What the asset must deliver, under what conditions, and what it depends on outside the main unit.
- Responsibilities. Owner, operator, technical authority, contractor and independent assurance, by role. Where one person in your company holds several, say so.
- Who owes the next decision. That person by role, and any chain of referrals that has taken its place.
- The decision record. Current disposition, its validity conditions, outstanding work and escalation trigger. If funding is unavailable, say so.
- Field label to board paper. Identity, boundary, records, and physical and contractual interfaces. State honestly where the chain breaks.
- The honest part. What your company currently gets wrong here. No AI in use yet is normal.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 2 — The required function, who owes the next decision, and the asset verified from field label to board paper.
How this is marked — the rubric
| States the required function with a system view naming dependencies outside the main unit | 20 |
| Separates the five responsibilities by role and names who owes the next decision | 25 |
| Drafts a decision record with disposition, validity conditions, outstanding work and an escalation trigger | 25 |
| Verifies identity, boundary, records and interfaces, and states honestly where the chain breaks | 30 |
| Total · pass mark 70 | 100 |
Submitting session 2
Write your answer in Word or Google Docs. The first line must read
Write your answer in Word or Google Docs. The first line must read
Session: 2, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.Capabilities Covered
Defining required functionSeparating responsibilitiesOne accountable decision recordVerifying asset identityMapping physical and contractual boundaries
Chapter 3 · The Foundations You Cannot SkipSESSION 3
Integrity Under Uncertainty
Why This Matters
A condition report rarely arrives with every answer you would like. Some locations may have been inaccessible; an earlier reading may have used a different method. Uncertainty becomes a management failure when it disappears from the decision pack, and a model can make that very easy by turning cautious prose into a smooth narrative. Today you learn to keep what was seen apart from what it means, to treat disagreement as information, and to judge what further evidence is worth before the chance to act closes. You leave with the evidence record for your own decision.
25-Minute Module
0:00 – 5:00
Four things that look alike on one screen
- Plain language first. Someone records what was seen; someone explains what it may mean; someone extends that into the future; someone authorises what may be done. Four acts that must stay distinguishable on one screen.
Layer What it is Example Observation What was recorded, where, when and how A measured thickness Interpretation What the observation may mean A corrosion rate inferred from comparable measurements Prediction The interpretation extended into a future condition A predicted remaining interval Acceptance decision An authorised disposition within specified limits An authorised inspection interval - THE DOCTRINE
A measured thickness is not a corrosion rate. A corrosion rate is not a guarantee of future deterioration. A predicted remaining interval is not an authorised inspection interval.
- You need not perform each calculation, but you must see why the distinctions affect confidence and authority. Verify any value extracted from a scan against the source: a decimal point, unit or location error can change the conclusion without making the prose look unusual.
When the layers blur, ask which one you are being asked to approve.
5:00 – 10:00
The evidence record, and disagreement as information
- For each material item, retain the asset identity, inspection location, date, method, instrument or procedure reference where relevant, raw result, assessor, report revision and acceptance status. Competent specialists select the precise fields. The principle: another qualified person should be able to understand what was observed and reproduce the reasoning to the extent the method allows.
- Record the source date separately from the upload date; a recent upload can contain old observations. Control which drawing version supports current decisions. Search ranking is not a technical authority: the first result should not become the accepted basis because it was easy to retrieve.
- Suppose two records appear to show an unexpected increase in wall thickness. Possible explanations include different measurement locations, different preparation, different methods, a repair, data-entry error or measurement uncertainty. None should be silently selected because it makes a trend easier to fit.
- Ask what evidence would resolve the disagreement and whether the unresolved state changes the operating decision: a targeted reinspection, or a competent reviewer reconciling the records. Use AI to assemble the conflicting records, flag differing units and revisions and prepare questions for the reviewer.
- THE DOCTRINE
Do not reward a system for producing one answer at all costs. A controlled refusal to draw a trend can be the correct output.
Accessibility is not validity.
10:00 – 17:00
A fictional teaching case with conflicting readings
- Every figure here is the book's assumption for teaching: not field data, not equipment life, not engineering guidance. Recompute it with me.
Step Assumed figure Earlier recorded thickness of a piping component 8.4 mm Recorded two years later 7.8 mm Simple difference, if locations and methods are comparable 8.4 − 7.8 = 0.6 mm over two years, or 0.3 mm per year Relevant minimum, if a competent assessment had established it 6.6 mm Arithmetic margin 7.8 − 6.6 = 1.2 mm Margin divided by the assumed rate 1.2 ÷ 0.3 = four years Third record, six months after the 7.8 mm reading 8.1 mm - Those four years are not a fitness-for-service result or a recommended inspection interval. It ignores reading uncertainty, possible localised damage, changes in service and the basis of the minimum, which this arithmetic cannot choose; it needs the applicable engineering assessment.
- Now the third record. A naive model might conclude that deterioration has slowed or reversed. The engineer instead asks whether the measurements refer to the same location and method. Until that is established, a more elaborate model may only be more elaborate speculation.
- AT THE ASSET IN THE FIELD · THE WORKED CASE
- In the field. The integrity engineer's question is comparability: same location, same method, same preparation? Until answered, the 8.1 mm reading is neither averaged in nor discarded, and the technical authority defines what may be done meanwhile.
IN THE BOARDROOM · THE PARALLEL- In the boardroom. The decision paper states which assumptions have evidence, which need verification and what action is appropriate meanwhile. The leader does not choose a rate; the leader makes sure the question reaches the right specialist with the resources and time to answer it.
A teaching calculation is useful precisely because it exposes its assumptions.
17:00 – 21:00
What more information is worth
- More information has value when it can change a decision before the opportunity to act closes. Ask what finding would cause the team to choose differently. If no plausible result would change the action, another study may add little. If one targeted examination could distinguish a local repair from a major replacement, it may be valuable.
- THE DOCTRINE
An information-value calculation is not permission to continue operating while evidence is inadequate. The technical authority defines what can be done during the investigation, including restrictions or shutdown where required.
- Set a deadline for resolving uncertainty. An item marked awaiting clarification for several reviews can become a de facto decision to continue. Make that decision explicit, with the appropriate authority, basis and limits.
- Measures for leadership: open evidence conflicts affecting material decisions, their age and whether interim dispositions are current. Test model outputs for source fidelity with reports that say cannot confirm, not inspected and subject to further assessment. The goal is dependable support for the actual decision, not an accuracy figure averaged across easy records.
Awaiting clarification is a decision; make someone with authority own it.
21:00 – 25:00
Write it down, Section 3, and Pop Quiz 1
- Before you close the laptop, write the one uncertainty that could change your decision, and the date after which waiting becomes a decision in its own right.
- THE DOCTRINE
Ask which uncertainty could change the decision. Fund its resolution and preserve it visibly until the qualified reviewer closes it.
- Section 3 of your plan is The evidence record, the uncertainty that could change the decision and its deadline.
- Pop Quiz 1 opens now: five questions, closed book, on Sessions 1 to 3. Sit it before you start Session 4.
Preserve the doubt until someone qualified closes it.
Key Terms
Observation and interpretation
An observation is what was recorded, where, when and how; an interpretation is what it may mean. Keep them visibly apart.
Acceptance decision
An authorised disposition within specified limits. A prediction or calculated interval does not become one until the appropriate authority accepts it.
Source fidelity
Whether a model's output preserves what the source record actually says, including negations, units and qualifiers such as not inspected or cannot confirm.
Value of information
Further evidence is worth obtaining when a plausible result could change the decision before the opportunity to act closes.
Self-Check — Before You Submit
1. In the book's fictional case, readings of 8.4 mm and then 7.8 mm two years later meet an assumed minimum of 6.6 mm. What does the arithmetic give?
- A. Two years of remaining life
- B. Four years: the reinspection interval
- C. Four years: a teaching calculation, not a fitness-for-service result or inspection interval
- D. Six years
C. Four years: a teaching calculation, not a fitness-for-service result or inspection interval — 0.6 mm over two years is 0.3 mm per year; the margin is 7.8 − 6.6 = 1.2 mm; 1.2 ÷ 0.3 is four years. Reading it as an inspection interval is the trap.
2. A third reading of 8.1 mm arrives six months later, and a vendor's model reports that deterioration has reversed. What should leadership require?
- A. Preserve all three readings and establish location and method comparability first
- B. Discard 8.1 mm as an outlier
- C. Accept the model, since it uses more data
- D. Average the three readings
A. Preserve all three readings and establish location and method comparability first — Discarding or averaging silently selects an explanation to fit a trend. Until comparability is established, a more elaborate model may only be more elaborate speculation.
3. A team argues that a further study is worthwhile, so the unit should run normally until it reports. What is wrong?
- A. Nothing, if the study is funded
- B. Studies cannot run during operation
- C. An AI prediction should replace the study
- D. Information value is not permission to operate on inadequate evidence; the technical authority sets the interim position
D. Information value is not permission to operate on inadequate evidence; the technical authority sets the interim position — Funding the study is the tempting answer but does not settle the interim position, which the technical authority sets, including restrictions or shutdown where required.
Today’s Assignment
Write the evidence record for your decision: what has been observed, interpreted, predicted and accepted, the uncertainty that could change it, and the deadline after which waiting becomes a decision.
- Four layers, kept apart. For your company's decision, what has been observed, what has been interpreted, what has been predicted and what has actually been accepted, and by whom.
- The evidence and its limits. For each material item: source, date, method, revision and acceptance status.
- The uncertainty that matters. The disagreement or unknown that could change the decision, and what would resolve it.
- Is it worth it? Whether resolving it is worth the cost before the opportunity to act closes, and what result would make you choose differently.
- The interim position and the deadline. What may be done meanwhile and who sets it, and the date after which awaiting clarification becomes an explicit decision with an authority, basis and limits.
- The honest part. Where your business has let uncertainty vanish from a decision pack.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 3 — The evidence record, the uncertainty that could change the decision and its deadline.
How this is marked — the rubric
| Separates observation, interpretation, prediction and acceptance for the student's own decision | 25 |
| Records evidence with source, date, method, revision and acceptance status, and states its limitations | 20 |
| Names the uncertainty that could change the decision, what would resolve it and whether that is worth the cost in time | 30 |
| States the interim position with who sets it and a dated deadline for an explicit decision | 25 |
| Total · pass mark 70 | 100 |
Submitting session 3
Write your answer in Word or Google Docs. The first line must read
Write your answer in Word or Google Docs. The first line must read
Session: 3, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.ASSESSMENT NOW OPEN
Pop Quiz 1 — Ownership, Identity and Uncertainty opens now. Sit it before you start the next session — it is how you find out whether the doctrine landed, and the practice papers may be resat as often as you like.
Sit Pop Quiz 1 →Closed book, no notes, and Coach Ritchie switches himself off while you sit it.Capabilities Covered
Separating observation from acceptanceBuilding an evidence recordTreating disagreement as informationValuing further evidenceSetting deadlines on uncertainty
Chapter 4 · The Foundations You Cannot SkipSESSION 4
Criticality Before Prediction
Why This Matters
Maintenance resources are finite. A list ranked only by likelihood of failure can direct them to the wrong work: a frequently failing item of limited consequence ahead of a rarely demanded protective function whose failure would matter greatly. Today you learn to read a risk score for what it is, to test a claim of redundancy, to keep exposure apart from readiness, and to stop necessary operating conditions competing with discretionary projects for the same budget. Your decision will take its honest place in your company's portfolio.
25-Minute Module
0:00 – 6:00
Function and consequence first, and what a score means
- Criticality begins with function and consequence. Ask what happens if the equipment cannot perform its required duty, what depends on it and what protection remains. Then examine condition and likelihood of failure. The result should support a decision about inspection, maintenance, contingency or investment, not a ranking for display.
- Plain language first. Many businesses rate likelihood and consequence as low, medium or high and number the categories. Those numbers record order, like places in a queue: third is behind second, but not by a measured distance. Multiplying them does not create a measured annual probability or an expected monetary loss. A score of twenty is not necessarily twice the risk of ten.
- THE DOCTRINE
Use categories to identify escalation and required review. Use a quantitative model only when its inputs and structure justify it. Where consequences include major safety or environmental harm, do not reduce the decision to one commercial number.
- Risk-based inspection has an established technical discipline. API's public catalogue distinguishes programme guidance in RP 580 from the methodology of RP 581, and separately describes fitness-for-service assessment. The book does not reproduce those methods, and neither does this session. Require a competent basis for the chosen approach and a clear account of what the ranking means.
Ask what the number means before you ask how high it is.
6:00 – 11:00
Dependencies, common causes and the claim of redundancy
- Two installed pumps do not automatically provide independent resilience. They may share a power source, suction condition, cooling service or maintenance weakness. The standby may never have been demonstrated under the required duty. A model that treats the units as independent because there are two tags can overstate availability.
- Ask what can disable both units, whether the standby can start and carry the load, and how quickly service can be restored, including the people and materials needed. A spare in a remote warehouse may not be available within the time a reliability calculation assumes.
- AT THE ASSET IN THE FIELD · THE WORKED CASE
- In the field. In a fictional gas plant, two compressor tags sit on the register. The reliability engineer walks the dependencies: shared power, shared cooling, one crew, one spare held far from site. The redundancy claim becomes a set of questions, not a number.
IN THE BOARDROOM · THE PARALLEL- In the boardroom. The same plant proposes improving one compressor. The benefit case must follow the actual system: if the common evacuation route remains constrained, the local improvement can produce little additional saleable volume, however well the compressor performs.
- Revisit criticality when service changes. A once nonessential component can become a bottleneck after a modification, and a new process configuration can introduce dependencies the old assessment did not consider. Keep the criticality basis tied to the actual plant.
Two tags are a claim of redundancy, not proof of it.
11:00 – 16:00
Separate exposure from readiness
- Two different questions. Exposure is how much it matters: it sets urgency and the escalation required. Readiness is whether the job can actually be done: scope, access, materials, competence and approvals in place. A high-priority item may not be ready, and a ready low-risk job should not make a critical unready job disappear from the management view.
- Give unready critical work an explicit route to readiness: the missing decision or resource, its owner and the required date. That lets leaders act on the constraint. A backlog dashboard that shows only planned completion dates may encourage planners to assign dates without a credible means of delivering them.
- THE DOCTRINE
AI can link work orders to inspection findings, material availability and access requirements, and propose a sequence for review. It should not downgrade an exposure because the job is difficult to execute or lacks funding.
- Measures for leadership: material exposures with no current disposition, safety-critical work overdue against its applicable requirement, and critical jobs held by an unresolved readiness constraint. Use counts and ages with context; one significant item can matter more than a large number of minor tasks.
Hard to do is never a reason to look less dangerous.
16:00 – 21:00
Prioritisation under a constrained budget
- Begin with necessary obligations and unacceptable conditions. These are gates, not optional projects competing on return. If the business cannot meet them, leadership must revisit the operating plan and the authority for continued service. A deficient condition cannot be made acceptable by assigning it a lower budget priority.
- For feasible discretionary options, compare the value protected, implementation cost, outage consequences, uncertainty and dependencies, and state whether benefits overlap. Two projects that both claim the same avoided downtime cannot simply have their business cases added together.
The book's fictional portfolio Its state Where it belongs A deteriorating transfer line An unresolved acceptance issue A gate: resolve its necessary conditions for operation first A recurring compressor fault A defined investigation and repair option Compared on actual business merits and funding constraints A warehouse automation proposal A positive estimated efficiency benefit Compared on actual business merits; a favourable return cannot authorise an unacceptable line condition - The human difficulty: the warehouse project may be easier to explain and deliver, while the line investigation exposes uncertainty nobody wants to own. Do not let presentational clarity decide. Use scenarios: if the replacement slips beyond the next access window, what changes?
- THE DOCTRINE
Keep necessary operating conditions outside the competition for discretionary returns. Then allocate scarce resources using a system view of consequence, readiness and value.
A gate is not a project; do not let it enter the auction.
21:00 – 25:00
Write it down, and Section 4
- Before you close the laptop, write whether your decision is a necessary gate or a discretionary choice, and the one dependency that could undo any redundancy your company has been relying on.
- If your decision is critical but not ready, name the missing decision or resource now, with its owner and required date.
- Section 4 of your plan is Criticality, dependencies, exposure kept apart from readiness, and the gates outside the budget contest.
Rank honestly, then fund the means of doing the work.
Key Terms
Ordinal risk score
A ranking built from categories such as low, medium and high. It supports escalation and review; it is not a probability or expected loss.
Common cause
A shared power source, service, maintenance weakness, crew or spare that can disable apparently redundant equipment at the same time.
Exposure and readiness
Exposure is how much a condition matters; readiness is whether scope, access, materials, competence and approvals exist. Report them separately.
Necessary gate
A legal condition, protection requirement or engineering acceptance that must be met before options are compared on return. It cannot be traded against commercial benefit.
Self-Check — Before You Submit
1. On your company's matrix, asset A scores 20 and asset B scores 10. What can you safely conclude?
- A. A carries twice B's expected loss
- B. A ranks higher for escalation and review; the score is not a probability
- C. A must be funded before B
- D. B can leave the register
B. A ranks higher for escalation and review; the score is not a probability — Multiplying ordinal categories does not create a probability or expected loss, so twenty is not necessarily twice ten. Treating the score as a funding order is the second trap.
2. Finance proposes one weighted score across a transfer line with an unresolved acceptance issue, a compressor fault and a warehouse automation proposal. What should you do?
- A. Accept the score if the weights are agreed
- B. Fund the warehouse first for its return
- C. Fund all three equally
- D. Resolve the line's necessary conditions first, then compare the other two on merit
D. Resolve the line's necessary conditions first, then compare the other two on merit — The line's issue is a gate, not a project competing on return. A favourable warehouse return cannot authorise an unacceptable line condition, however tidy the agreed weights look.
3. Planners propose lowering the priority of a critical job because materials are unavailable and access is not arranged. What is the right response?
- A. Keep its exposure and give it a route to readiness: missing resource, owner, date
- B. Accept the change as realistic
- C. Assign a completion date to close it
- D. Let the AI scheduler decide
A. Keep its exposure and give it a route to readiness: missing resource, owner, date — Exposure and readiness are separate; difficulty or lack of funding must not downgrade an exposure. A date without a credible means of delivery hides the constraint.
Today’s Assignment
Assess the criticality of the asset in your decision, test its dependencies, keep its exposure apart from its readiness, and show whether it is a necessary gate or a discretionary choice.
- Function and consequence. What the asset does and what happens to your company if it cannot perform its required duty, including what protection remains.
- Dependencies and common causes. What could disable it and any backup together: power, services, people, spares and their location, and outside parties. Say whether any claimed redundancy has been demonstrated.
- Exposure and readiness, separately. How urgent it is, and whether scope, access, materials, competence and approvals are in place. If critical but unready, the route to readiness: missing decision or resource, owner and date.
- Gate or choice. Which conditions are necessary gates outside the competition for discretionary funds, and how any remaining option would be compared with others without overlapping benefits.
- The honest part. Where your business has let a risk score, a funding gap or an easy project rank a necessary condition. No AI in use yet is normal.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 4 — Criticality, dependencies, exposure kept apart from readiness, and the gates outside the budget contest.
How this is marked — the rubric
| States function and consequence, and explains what the company's risk score does and does not mean | 20 |
| Tests claimed redundancy against named dependencies and common causes, including people and spares | 25 |
| Reports exposure separately from readiness and gives unready critical work a route with owner and date | 30 |
| Separates necessary gates from discretionary options and avoids counting overlapping benefits | 25 |
| Total · pass mark 70 | 100 |
Submitting session 4
Write your answer in Word or Google Docs. The first line must read
Write your answer in Word or Google Docs. The first line must read
Session: 4, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.Capabilities Covered
Reading risk scores correctlyTesting redundancy claimsSeparating exposure from readinessKeeping gates out of the budget contestSystem-view prioritisation
Chapter 5 · The Foundations You Cannot SkipSESSION 5
Inspection Must Resolve a Decision
Why This Matters
An inspection campaign can be on time, on budget and full of photographs, and still leave the central question unanswered. Coverage may be generous where access is easy and thin where a significant mechanism is hard to examine. A report can be issued while the acceptance decision stays open. Today you learn to commission inspection as an owner should: start from the decision it must resolve, match the method to the question, agree the evidence before mobilisation and follow the condition to an accepted disposition. You also learn to test a new technology without letting a demonstration stand in for qualification.
25-Minute Module
0:00 – 5:00
Start with the decision, then match the method
- Start with the purpose. Is the inspection meant to establish a baseline, characterise an anomaly, investigate a suspected mechanism, verify a repair or support a future operating period? The answer affects method, coverage, timing and the competence required to interpret the result.
- Different methods see different things. A visual survey and a thickness examination answer different questions. External surface condition does not establish the condition of every internal location. A broad screening method may identify areas requiring detailed examination without supporting final acceptance on its own.
- Ask the inspection specialist what the chosen method can detect, its limitations and the consequence of incomplete coverage, in terms you can follow without reducing the procedure to a purchasing description. Procurement needs enough clarity to compare equivalent scopes; operations needs enough to prepare access and isolation.
- THE DOCTRINE
Agree the evidence deliverable before mobilisation: identifiers, units, revision control and exception reporting. A report that cannot link results to stable locations creates avoidable problems for the next inspection.
Name the question before you name the method.
5:00 – 11:00
Robotics changes access, not the acceptance obligation
- Robots, drones and remotely operated systems may offer useful ways to collect evidence or perform work. Suitability depends on the task and environment: reach, stability, navigation, surface conditions, inspection payload, recovery arrangements and the quality of location references. A demonstration on an accessible surface does not establish suitability for every offshore or process environment.
- NUPRC's 2026 Guidelines for the Adaptation of Novel Technology distinguish validation or verification, pilot testing and full-scale deployment, with approvals bounded by their stated scope. For an upstream owner, establish the applicable qualification and approval route before deployment. Other facilities determine their own relevant requirements rather than borrowing an upstream approval as a universal authorisation.
- THE DOCTRINE
Separate qualification of a technology from acceptance of an individual job. Even an accepted technology can produce unusable evidence if access, calibration, coverage or traceability are inadequate on the day.
- AT THE ASSET IN THE FIELD · THE WORKED CASE
- In the field. In a fictional shallow-water operation, a remotely operated system is mobilised for an external survey. On the day, location references are poor and part of the scope cannot be reached. The inspection lead records the gaps rather than reporting the campaign complete.
IN THE BOARDROOM · THE PARALLEL- In the boardroom. The same campaign reaches the investment committee as a request to extend the technology. The question is not whether it was fast, but whether its evidence supports the same decision as the alternative, judged by accepted coverage and usable findings.
A good demonstration proves a demonstration.
11:00 – 15:00
Judge evidence quality alongside productivity
- A contractor may offer faster coverage. Ask whether the resulting evidence supports the same decision as the proposed alternative. Compare accepted coverage and usable findings, not square metres traversed, images captured or hours saved. The commercial basis should make quality failures and repeat work visible.
- For image-based AI, use a reviewed test set that reflects the equipment, coatings, lighting and defect types within scope, including acceptable surfaces and difficult negatives. A system can perform well on obvious damage and struggle with the conditions that matter most to the inspection team.
- Keep localisation and classification distinct. Finding a visual feature is not sizing it, identifying its mechanism or establishing its acceptability. A model's confidence score does not supply missing engineering evidence; the specialist defines what further examination a flag requires.
Pay for evidence you can accept, not area you can invoice.
15:00 – 21:00
Completion includes disposition, and a useful pilot
- Use a chain of states that reflects the work: planned, access ready, examined, report checked, anomaly assessed, disposition accepted and follow-up entered. A job may reach one state without reaching the next. Do not use inspection complete as shorthand for risk resolved.
- Define the escalation route for an unexpected condition before mobilisation: who can assess it, what immediate controls may be needed and how to obtain an urgent decision. An offshore campaign is a poor time to discover that the only available reviewer lacks authority for the equipment in question. If further work is needed, retain the original anomaly identifier and link later records to it.
Pilot element What the book asks for Population Bounded, with a specific evidence problem Comparator A credible existing approach Measures Accepted coverage, reproducibility of location, source fidelity, review time and handling of inaccessible areas Full cost The time and cost of follow-up examination and rejection Review Blind where practical: qualified reviewers not told which method is expected to win; disagreements preserved Dependence Recorded in the support plan if the approach only works when one individual interprets every result - THE DOCTRINE
At the end of a pilot, ask whether it improved a real decision. Faster data collection can be valuable, but it is not an integrity improvement unless the evidence supports that claim.
Keep the anomaly's identity from first finding to final acceptance.
21:00 – 25:00
Write it down, Section 5, and Practice Exam 1
- Before you close the laptop, write the decision your next inspection must resolve, and the state in the chain where your company's last campaign on this asset actually stopped.
- THE DOCTRINE
Before approving the next inspection campaign, identify the decisions it must resolve and the evidence required for each. Pay for accepted evidence and an accountable disposition.
- Section 5 of your plan is The inspection the decision needs: method, evidence deliverable and the chain to disposition. With it, Week 1 of your plan is complete.
- Practice Exam 1 opens now: fifteen questions, closed book, on Sessions 1 to 5, closing Week 1. Sit it before you start Session 6.
An inspection is finished when the decision is, not when the report is.
Key Terms
Evidence deliverable
The form in which inspection results must be handed over, agreed before mobilisation: identifiers, units, revision control and exception reporting, linked to stable locations.
Accepted coverage
The part of the intended scope that produced evidence qualified reviewers accept for the decision, as distinct from area traversed or images captured.
Qualification and job acceptance
Qualifying a technology establishes that it can be used within a stated scope; accepting a job establishes that this campaign's evidence is usable. One does not imply the other.
Disposition
The accepted decision on what happens to an assessed anomaly. Inspection is not complete in the owner's sense until disposition is accepted and follow-up entered.
Self-Check — Before You Submit
1. A supplier demonstrates a drone inspection on an accessible onshore tank shell and proposes using it across your offshore facilities next quarter. What should the owner do first?
- A. Approve it, since the demonstration succeeded
- B. Approve it if the supplier holds an upstream approval for another operator
- C. Establish suitability for the actual task and environment and the applicable qualification route, then design a bounded pilot
- D. Reject robotic inspection as unproven
C. Establish suitability for the actual task and environment and the applicable qualification route, then design a bounded pilot — A demonstration on an accessible surface does not establish suitability offshore, and an approval is bounded by its stated scope. Outright rejection is the opposite error: such systems may be useful once qualified and tested.
2. Which statement is consistent with the chain of inspection states?
- A. Once the report is issued, the risk is resolved
- B. A job can be examined and its report checked while no disposition has been accepted
- C. Follow-up is entered before the anomaly is assessed
- D. Access readiness is the contractor's concern only
B. A job can be examined and its report checked while no disposition has been accepted — The seven states are distinct, and a job can reach one without the next. Treating report issue as resolution is the trap: inspection complete is not shorthand for risk resolved.
3. An image model flags a feature on a pipe support with a high confidence score. What does the flag establish?
- A. That the defect is sized and acceptable
- B. That no further examination is needed
- C. That the support should be replaced
- D. That a visual feature has been located; the specialist defines what further examination it requires
D. That a visual feature has been located; the specialist defines what further examination it requires — Localisation is not sizing, mechanism or acceptability. A confidence score does not supply missing engineering evidence, however high it is; the specialist decides what the flag requires.
Today’s Assignment
Specify the inspection your decision needs: the decision it must resolve, the method and its limits, the evidence deliverable, the chain to disposition and, if a new technology is proposed, a pilot design.
- The decision it must resolve. Baseline, anomaly characterisation, suspected mechanism, repair verification or a future operating period, and why.
- Method matched to the question. What the method your company would use can detect, its limitations and the consequence of incomplete coverage.
- The evidence deliverable. Identifiers, units, revision control and exception reporting, agreed before mobilisation.
- The chain to disposition. From planned to follow-up entered, with the escalation route for an unexpected condition and who holds authority. Say where your company's last campaign stopped.
- A pilot, if proposed. If a new technology or AI aid is in front of you: the bounded population, credible comparator, accepted coverage measure and blind review. If none is proposed, say so; no AI in use yet is normal.
- The honest part. Where your business has paid for activity rather than accepted evidence.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 5 — The inspection the decision needs: method, evidence deliverable and the chain to disposition.
How this is marked — the rubric
| Names the decision the inspection must resolve and matches the method to it with stated limitations | 25 |
| Specifies an evidence deliverable with identifiers, units, revision control and exception reporting | 20 |
| Maps the chain of states to accepted disposition with an escalation route and authority | 30 |
| Designs a pilot with population, comparator, accepted coverage and blind review, or justifies none, and names an honest past weakness | 25 |
| Total · pass mark 70 | 100 |
Submitting session 5
Write your answer in Word or Google Docs. The first line must read
Write your answer in Word or Google Docs. The first line must read
Session: 5, or the submission is refused. Upload the Word file (.docx), or switch to the Google Doc tab and paste your document’s share link.Marked against the rubric above and returned as a written letter, saved to your progress page. Do not submit until you have passed the self-check above.ASSESSMENT NOW OPEN
Practice Exam 1 — The Ownership Mandate and the Foundations opens now. Sit it before you start the next session — it is how you find out whether the doctrine landed, and the practice papers may be resat as often as you like.
Sit Practice Exam 1 →Closed book, no notes, and Coach Ritchie switches himself off while you sit it.Capabilities Covered
Matching method to decisionSpecifying evidence deliverablesQualification versus job acceptanceFollowing anomalies to dispositionDesigning a fair inspection pilot