30-Session Programme · 25 Minutes Per Day · Free · For leaders in Nigeria’s oil and gas sector
WEEK 1 OF 6 · LEADERSHIP AND THE OPERATING FOUNDATION · SESSIONS 1–5 · WORK AT YOUR OWN PACE
Orientation · The Responsibility to ActSESSION 1
The Responsibility to Act and the Reader's Toolkit
Why This Matters
A production shortfall can survive several meetings without becoming a decision. Everyone understands the problem and the barrels are still being lost. Over thirty sessions you will build a signed 90-Day Intelligent Operating Plan for your own company, one section per session, around one decision you choose today: a decision that currently costs your company time, money or operating opportunity. You will trace it from the first observation to the last required approval and action, and find where it stops. A model can improve the diagnosis. It cannot carry the decision through. That part is leadership, and it is what this programme trains.
25-Minute Module
0:00 – 5:00
Where action stopped, and the finish line
- The field report records a shortfall; the engineer explains it; maintenance names the part; procurement confirms the lead time; finance asks about the budget. By the next review the barrels are still being lost.
- THE DOCTRINE
A model may improve the diagnosis or provide an earlier warning. It cannot settle who may approve the purchase, make cash available or provide a safe work window. Those responsibilities sit with leadership.
- The six weeks: 1 Leadership and the Operating Foundation; 2 Governance, People and the Subsurface; 3 Wells, Production and Processing; 4 Reliability, Projects and Delivery; 5 Protecting Assets and Operating Relationships; 6 Commercial Value, Governance and Execution.
- The finish line: a signed 90-Day Intelligent Operating Plan for your own company, twenty-nine sections written one per session and signed in Session 30. House rules: write about your own company; bring a real decision; no value claim without a baseline; no decision without a named owner; report a blocker the same day. The pass mark is 70.
- Said once: this is leadership training, not investment, legal or engineering advice. The frameworks are the author's synthesis, not an official regulatory standard or a university-endorsed method. Consequential control, safety and legal matters need competent technical, operating and governance assurance.
Ask where action stopped before you ask what to buy.
5:00 – 10:00
From a barrel to an operator: five connected responsibilities
- This book applies the doctrine of The Intelligent Barrel across an integrated Nigerian operator: wells and facilities, and also the contracts, people, relationships and funding the plan depends on. Local success can hide system loss. A well can produce more fluid while adding little saleable oil. A production record can be accurate and still say little about the cash available for the next intervention.
Responsibility Who carries it 1. Establish a dependable operating foundation Leadership 2. Turn resource knowledge into feasible production choices Subsurface and well teams 3. Sustain usable capacity Processing, reliability and projects 4. Protect assets and maintain obligations and relationships The operator as a whole 5. Convert the operating plan into durable value and accountable execution Commercial and financial leadership - The sequence is a reading structure; in a real operator the five act together.
- On evidence: the IEA identifies AI applications across energy operations, including resource evaluation, production, maintenance and emissions work (IEA, Energy and AI, 2025). That is sector context. It does not establish the return an individual Nigerian asset should expect, and no universal percentage improvement is promised.
The connections deserve as much attention as the individual departments.
10:00 – 17:00
The six frameworks, as questions
Framework What the book says The question it forces 1. Operating-Intelligence Loop Sense, decide, act, learn. A dashboard mainly supports sensing; none completes the loop alone. Where does our decision stop? 2. Data-Foundation Precondition Evidence must be adequate for its intended use; adequacy depends on consequence. Is this evidence adequate for this decision? 3. Autonomy Ladder Retrieval and drafting; reviewed recommendation; bounded execution with explicit authority, monitoring and recovery. What permission has this use earned, and who can stop it? 4. Value-versus-Risk Map Benefit and failure consequence together. Some safety, legal, technical and authority conditions are necessary requirements. What does a wrong answer cost? 5. Sequencing Discipline Bounded problem; record and process; compare with simpler alternatives; test; expand only on evidence. Have we tried the simpler way? 6. Operations-Excellence Lens Judge it by the operating outcome it serves: production, reliability, quality, environment, cost, cash or assurance. Which outcome moves, against what comparison? - THE DOCTRINE
Permission belongs to the approved use, not to the model's general reputation. A stronger expected return does not cancel a failed necessary condition.
- A change of user, asset, input, service, interface or task can require reassessment. Record who can stop the application.
- A faster report does not prove lower deferment; an automated invoice does not prove collection. Keep the technical measure and the business outcome together; they answer different questions.
Use the frameworks as questions for a live decision, not as vocabulary.
17:00 – 22:00
Three closing disciplines, and your decision
- Know which method you are using. Instrumentation, historian, workflow automation, physics-based model, mathematical optimisation, machine learning and generative AI do different things. Calling it all AI hides different failure modes. Select the least complex credible method.
- Carry one decision record that distinguishes recommendation, approval, execution and assurance; where one person holds several roles, make that visible. Ask three questions at the end: what improved in the decision, what happened after approval, and what evidence supports the outcome.
- THE DOCTRINE
If the evidence is adequate but approval is unavailable, the next investment may concern authority. If approval exists but a spare cannot arrive in time, the next decision concerns supply. If no one verifies the effect, the benefits report remains uncertain.
- AT THE ASSET · THE WORKED CASE
- Asset. A fictional swamp operator proposes a workover on a declining well. The diagnosis is adequate, but nobody holds approval for the spend. The loop breaks at authority.
IN THE OPERATING REVIEW · THE PARALLEL- Review. The same company's monthly review approves another workover, the work is done, and nobody checks the result against a comparison. That loop breaks at learn. One loop, two ends, different failures.
- Now choose your company's role (operator, partner, service company or other) and one decision that currently costs your company time, money or operating opportunity.
Repairing a measurement or removing an approval delay can be the right outcome.
22:00 – 25:00
Write it down, and Section 1
- Before you close the laptop, write three lines: your company and its role; the one decision; the stage where its loop breaks.
- Section 1 of your plan is Our company, the one decision that costs us most, and where its loop breaks. Choose a decision you will still care about in Session 30.
- Be willing to find that AI is unnecessary here. The programme succeeds when the business makes better decisions, even if fewer models are commissioned.
A recommendation needs evidence. An action needs an owner. An owner needs authority and resources.
Key Terms
Operating-Intelligence Loop
Sense, decide, act and learn: observe, choose an authorised response, act through a named owner, check the outcome.
Autonomy Ladder
Conceptual permission levels: retrieval and drafting, reviewed recommendation, then bounded execution with explicit authority, monitoring and recovery.
Decision record
A concise record of the choice, alternatives, evidence, uncertainty, constraints, owner, resources, expected consequence and review date.
Baseline
The agreed comparison used to evaluate a change. Its period, conditions and definition should be recorded.
Self-Check — Before You Submit
1. At a fictional onshore operator, a gathering-line problem is flagged early, confirmed and approved for repair, but the spare cannot arrive in time. What does the next decision concern?
- A. A more accurate predictive model
- B. More sensors on the gathering line
- C. Supply
- D. A new executive dashboard
C. Supply — The book's own test: approval exists but the spare cannot arrive, so the decision concerns supply. A better model is tempting, but sensing and deciding already work here.
2. A forecasting model has performed reliably on one asset. A team proposes using it, with the same permission, for a new task on another asset. What does the Autonomy Ladder say?
- A. It transfers, because the model has earned trust
- B. Permission belongs to the approved use; a new asset or task can require reassessment
- C. It may proceed once the vendor certifies the model
- D. The model must be recertified on the ladder
B. Permission belongs to the approved use; a new asset or task can require reassessment — Permission attaches to the approved use, not the model's reputation. Recertification sounds rigorous, but the ladder is conceptual permission levels, not a certification scheme.
3. A team says its application produces the daily report faster and claims this has reduced deferment. How does the Operations-Excellence Lens apply?
- A. Accept it; faster reporting must help
- B. Withdraw the application; speed is not an outcome
- C. Count daily users of the report
- D. Ask for the connection to deferment and a comparison, with attribution agreed by finance and the operating owner
D. Ask for the connection to deferment and a comparison, with attribution agreed by finance and the operating owner — A faster report does not prove lower deferment, but it may still be useful, so withdrawing it is also wrong. The lens asks for the connection and the comparison.
Today’s Assignment
Name your company, its role and one decision that costs it time, money or operating opportunity; trace that decision to its last required approval and action; and name where its loop breaks.
- Your company and its role. What it does, its scale and role. A partner or service company may choose the operator decision it supports.
- The one decision. What is decided, how often, and what delay costs. Write not known rather than invent a figure.
- The trace. From first observation to last required approval and action: who observes, judges, approves, funds, acts and checks. Mark where it waits.
- The break. The loop stage where it stops, the evidence, and the next investment it implies: evidence, authority, supply, funding or verification.
- The honest part. What your company currently gets wrong about this decision. No AI in use yet is normal; work from the decision in front of you. The answer may be a simpler method or no model.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 1 — Our company, the one decision that costs us most, and where its loop breaks.
How this is marked — the rubric
| Names the company, its role and one specific decision, its cost stated or marked not known | 25 |
| Traces first observation to last approval and action, naming roles not departments | 30 |
| Names the breaking loop stage with evidence and the next investment it implies | 30 |
| The honest paragraph names a failing in the student's own company, not the industry | 15 |
| 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
Locating where action stopsThe six toolkit frameworksThe three closing disciplinesDecision selectionPlan authorship
Part One · Leadership and the Operating FoundationSESSION 2
The Leadership Mandate
Why This Matters
A stalled operating decision is not always short of analysis. This chapter opens with an expenditure everyone agrees is necessary and nobody can approve. Better prediction might buy more time, but time is useful only if your company can use it. This session turns the decision you chose in Session 1 into a mandate: written in ordinary operating language, with an owner, an authorised route to action, the money to respond and evidence a board can challenge, including the cases that went badly. If that is uncomfortable to write, the route has probably never been mapped. Better to find out now.
25-Minute Module
0:00 – 5:00
The expenditure nobody can approve
- An illustrative operating situation, not a reported incident: a compressor begins to show an abnormal condition. The engineer wants an inspection and a spare assembly. Maintenance has a feasible work window. Procurement needs a purchase approval.
- The asset manager believes the expenditure belongs to a central budget; the central team believes it belongs to the asset. An analytical system sends increasingly confident warnings. None of them changes the approval arrangement. The equipment continues to deteriorate because the decision has no effective owner.
- The book's diagnosis is short: the operator has an authority problem. Better prediction might give it more time, but time is useful only if the organisation can use it.
- THE DOCTRINE
An early-warning application without a credible intervention budget can become a better description of deterioration.
Before you fund a sharper warning, check who can act on the current one.
5:00 – 11:00
What leadership must actually own
- A board should expect management to name the operating problem, the decision that will change, the person responsible, the resources required and the evidence that will establish value. A proposal naming only a platform and a vendor is incomplete.
- The chief executive does not need to select every algorithm. The chief executive does need to decide which operating outcomes deserve attention, how conflicting priorities will be resolved and who may commit resources. An asset manager cannot resolve a treasury constraint alone.
- I recommend a small initial mandate: a few recurring decisions where the company already understands the consequence of delay. For an integrated operator the book suggests production deferment, critical maintenance readiness and the reconciliation between delivered product and collected cash. Choose from evidence, not from the visual appeal of a demonstration.
- THE DOCTRINE
Write the mandate in ordinary operating language: improve the identification and resolution of recurring compressor-related deferment, using validated evidence and existing technical authority.
- The word improve still needs a baseline and acceptance criteria. Even so, that sentence is closer to an executable assignment than a general ambition to become AI driven.
If your mandate could be pasted into another company's strategy, it is not yet a mandate.
11:00 – 17:00
Authority before enthusiasm
- An accountable owner needs an authorised route to action, which may involve engineering approval, procurement, budget release, partner consent and a work permit. Map it before the pilot begins. Where several approvals are needed, identify the sequence and the person responsible for resolving disagreement.
- Delegation has limits: a bounded repair within budget may sit with the asset manager while a replacement goes to the investment committee. The system may assemble the evidence and prepare the request; it does not acquire purchasing authority because the recommendation is persuasive.
- Define what happens when information conflicts. If the production system indicates a loss while the maintenance record shows the equipment available, the useful response is a reconciliation with a named technical reviewer. It is not a contest in which the most confident dashboard wins.
- Money as well as signatures. Finance should review the practical funding requirement, including the period before benefits appear. The business case must include the cost of responding to an alert, not merely the cost of generating it.
- AT THE ASSET · THE WORKED CASE
- Asset. At a fictional flow station a pump warning is credible and maintenance has a work window. The bounded repair sits within the asset manager's delegated budget, so the route is short: engineering approval, purchase, permit, work.
IN THE OPERATING REVIEW · THE PARALLEL- Review. The same pump's full replacement exceeds that delegation and goes to the investment committee. The evidence is identical; the route, the approver and the funding are not. The mandate should name both routes before the warnings start.
Map the signatures and the money before the first alert, not after the tenth.
17:00 – 22:00
Evidence a board can challenge, and the operating conversation
- Software deployment is an activity. Operating improvement is an outcome. The board should see the difference in the reporting it receives.
Evidence the board should see What it reveals Change in decision time Whether the route to action got shorter Proportion of accepted alerts followed by completed actions Whether warnings become work Cost of those actions What responding actually costs Observed operating result Whether the condition changed Adverse evidence: false alarms, unnecessary interventions, missed events, reasonable rejections Whether expansion is justified - Before introducing the application, record the present method: how the problem is recognised, how long verification takes, where requests wait. That is your comparison, and it often reveals an improvement that needs no AI. Keep the option of doing nothing or a simpler method, without treating it as a personal defeat.
- THE DOCTRINE
Usage can indicate adoption, but repeated use of an unhelpful system is not value.
- The sponsor asks: which loss recurs, what evidence supports the cause, who verified it, what can be done within present authority, what needs further approval, and when the outcome will be reviewed. I would keep the first executive report short: decision, evidence, resources, result.
Reporting only the favourable cases deprives leadership of what it needs to judge expansion.
22:00 – 25:00
Write it down, and Section 2
- At the end of the first review cycle the book returns to the compressor. Did the company authorise the spare, was it available, was the work carried out safely, and did the intervention change the operating condition. If the answer stops at a recommendation, the leadership task remains unfinished.
- Before you close the laptop, write one sentence of mandate for your Session 1 decision and list every signature between the first observation and the action.
- Section 2 of your plan is The mandate: owner, authority route, funding and the evidence the board can challenge.
The summary should make challenge easier, not hide uncertainty under polished language.
Key Terms
Mandate
A written instruction in ordinary operating language naming the operating problem, the decision that will change, the person responsible, the resources and the evidence that will establish value.
Authorised route to action
The sequence of approvals a decision needs before work can happen, such as engineering approval, procurement, budget release, partner consent and a work permit, with a named person to resolve disagreement.
Adverse evidence
False alarms, unnecessary interventions, missed events and occasions when users reasonably rejected a recommendation. Without it, leadership cannot judge expansion.
Technical authority
A person or function with the defined competence and responsibility to accept the technical basis.
Self-Check — Before You Submit
1. Compressor warnings at a fictional operator grow more confident weekly, while the asset and central teams each say the spare belongs in the other's budget. What should the sponsor do first?
- A. Commission a more accurate predictive model
- B. Settle who owns the expenditure and map the authorised route to action, including funding
- C. Let the analytical system raise the purchase order, since its confidence is high
- D. Wait for the next review cycle to see whether the warnings persist
B. Settle who owns the expenditure and map the authorised route to action, including funding — This is an authority problem, not a prediction problem. Letting the system raise the order is tempting, but a persuasive recommendation does not acquire purchasing authority.
2. Which report best supports a board deciding whether to expand an early-warning application?
- A. The number of users who opened the application each week
- B. The number of dashboards deployed across assets
- C. Completed actions per accepted alert, their cost, the observed result, and the false alarms and reasonable rejections
- D. Vendor case studies from other operators
C. Completed actions per accepted alert, their cost, the observed result, and the false alarms and reasonable rejections — The board needs outcome evidence including adverse evidence. Usage is the tempting measure, but repeated use of an unhelpful system is not value.
3. The production system shows a loss while the maintenance record shows the compressor available. What is the right response?
- A. Trust the production system, because it measures flow directly
- B. Accept whichever dashboard reports higher confidence
- C. Refer the conflict to the software vendor
- D. Reconcile the records with a named technical reviewer
D. Reconcile the records with a named technical reviewer — Conflicting information calls for reconciliation with a named technical reviewer. Trusting production data looks sensible, but it still picks a winner without establishing which record is right.
Today’s Assignment
Write the mandate for your decision, map its authorised route to action and funding in your company, and name the evidence, including adverse evidence, the board should see.
- The mandate. One or two sentences in ordinary operating language. State what improve is measured against.
- The owner and the route. The accountable owner by role and every approval between recommendation and action in your company. Name the delegation limit and who resolves disagreement.
- The funding. Who pays for the response, from which budget, and the period before any benefit appears. Include the cost of responding, not only of detecting.
- The evidence for the board. The present method as comparison, the measures you will report, and the adverse evidence you will show. Keep doing nothing or a simpler method as options.
- The honest part. Where your company's route currently breaks: the signature nobody holds, the budget nobody owns or the bad news nobody reports. No AI in use yet is normal.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 2 — The mandate: owner, authority route, funding and the evidence the board can challenge.
How this is marked — the rubric
| States a mandate in ordinary operating language tied to the Session 1 decision, with its comparison named | 20 |
| Maps the complete authority route with an owner by role, the delegation limit and who resolves disagreement | 30 |
| Shows funding for the response, including the cost of responding and the period before benefit | 20 |
| Specifies board evidence, including the present-method comparison and named adverse evidence | 20 |
| The honest paragraph names where the student's own company's route currently breaks | 10 |
| 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
Writing an executable mandateMapping authority and delegationFunding the responseBoard-grade evidenceRunning the operating conversation
Part One · Leadership and the Operating FoundationSESSION 3
Managing the Operator as One System
Why This Matters
Your departments can all report good results while your company loses value. Wells deliver more fluid, processing holds throughput, the refinery builds inventory, commercial records orders and treasury stays short of cash. Each statement can be true. This session asks you to stop reading the business by department and draw the actual route from the reservoir, or your first activity, to payment for the decision you have chosen. On that route one condition limits what the whole system can deliver. Find it, write it into a single register with evidence and an owner, and you will know whether the improvement you are planning would reach cash or simply move the queue.
25-Minute Module
0:00 – 5:00
Draw the route to payment
- Those statements describe different parts of the chain, often on different time horizons. Leadership must see how they connect.
- The first task is to draw the actual route from the reservoir to payment. Include the equipment, storage, transport, contractual handoffs and approvals on which that route depends.
- An organisational chart is not an adequate substitute. It describes reporting lines; it does not show where a barrel waits or where a commitment becomes cash exposure.
- If your company is a service company or a non-operated partner, draw the route of the operator decision you support, from your first activity to the point where your company is paid.
Follow the barrel and the money, not the reporting lines.
5:00 – 11:00
Finding the binding constraint
- In plain terms, think of the narrowest point in a line: widening it anywhere else changes nothing. A constraint limits the performance of the system under the conditions being examined. It may be a physical capacity, a safe operating limit, a customer specification, a permit, a transport window or a funding restriction. Its location can change.
- In the book's illustrative week, a gathering line limits well delivery. After repair, separation capacity becomes the limit. After a processing improvement, export availability restricts saleable output.
- THE DOCTRINE
Additional well capability is valuable only to the extent that the remainder of the chain can accommodate it.
- So the management question is not how to increase production. It is how to improve feasible, saleable output and its economic contribution. The answer may be a well intervention, a meter repair, a compressor overhaul, an evacuation agreement or a revised delivery schedule.
- A single output labelled optimal can conceal which prices were used, whether losses and quality adjustments were included, whether the transport route is actually available and whether it fits the approved operating envelope. The model should expose those assumptions to the people who own them.
Ask what binds next before you pay to remove what binds now.
11:00 – 16:00
Oil, gas and refining in one decision, operated and non-operated
- Sending additional crude to refining may improve utilisation while reducing export volumes. Compare the full contribution from each feasible route: conversion costs, product yields, logistics, quality, collection timing and relevant obligations. A headline product price is insufficient.
- Gas adds a dependency. Compression availability may affect gas sales, gas lift, oil production and flaring. Reconcile shared costs and overlapping benefits rather than allocating them to whichever department presents most strongly.
- AT THE ASSET · THE WORKED CASE
- Asset. At a fictional integrated asset, a compressor repair would support gas sales and gas lift and reduce flaring. Each benefit is real at the plant, and each department wants to count it.
IN THE OPERATING REVIEW · THE PARALLEL- Review. At the monthly review the repair appears in three departmental cases. The review reconciles them into one figure agreed by technical, commercial and finance before ranking it.
- For operated assets, management directs work within its delegated authority. For non-operated interests, it exercises contractual rights through partner governance, technical assurance and approval processes. An AI recommendation does not change that relationship.
- The information model should distinguish gross asset output, the company's entitlement and the cash received, and an operational recommendation from a partner-approved commitment. Those distinctions belong in the decision record, not in a footnote added after a discrepancy is found.
No universal ranking rule fits every configuration; make the relationships inspectable.
16:00 – 21:00
One register, one scorecard, and a Friday meeting
Constraint register field What to write Asset or handoff Where in the route the restriction sits Observed restriction What is limited, and how Evidence The record that shows it Operating consequence What the system loses while it binds Owner A named role Proposed resolution and approval required What would remove it, and who must approve Time basis Tomorrow's scheduling issue or next year's capacity limit - Keep it selective, and preserve history: when a restriction is removed, record what became limiting next. That shows whether investment improves the system or simply moves the queue.
- The executive scorecard should include feasible saleable output, delivery reliability, reconciled inventory, realised contribution, cash collection and the age of unresolved constraints. Explain the difference between capability and realised value.
- The book's fictional Friday meeting: production can supply another 1,000 barrels on Saturday. The refinery can accept 600, export 500 and storage can temporarily hold 300. On paper those total 1,400, but they do not add to an instruction to produce 1,400 barrels. The team must establish which commitments share tanks, pumps and the same transfer window.
- Commercial reports a stronger refinery contribution, but the buyer collects only on Monday and product storage is nearly full; export offers a lower contribution with a confirmed movement; treasury needs receipts within the following week. The team prepares two feasible schedules for the same 1,000 barrels, with transfers, daily inventory, costs and receipt dates, plus the option of producing less. The COO approves the integrated plan, records why it differs from each department's preference, and checks actuals on Monday.
- THE DOCTRINE
Allocation decisions need a time dimension and a shared physical basis.
A constraint carried forward out of habit may no longer be binding; examine it before you report it again.
21:00 – 25:00
Write it down, Section 3, and Pop Quiz 1
- Before you close the laptop, sketch the route for your decision on one page and circle the point where it waits longest.
- Section 3 of your plan is The route to payment, the binding constraint and our constraint register.
- Pop Quiz 1 opens now — five questions, closed book, on Sessions 1 to 3. Sit it before you start Session 4.
The sophistication of the models matters less than the coherence of the decisions they support.
Key Terms
Binding constraint
The condition that limits the system's feasible performance for the period and decision being examined. Its location can change.
Constraint register
A selective common record of the most material constraints: asset or handoff, restriction, evidence, consequence, owner, resolution, approval required and time basis, with history preserved.
Operating envelope
The approved conditions and limits within which an asset or activity may operate, established through the applicable competent process.
Gross output, entitlement and cash
Three different quantities: what the asset produces, the company's share of it, and the money actually received. The decision record should keep them distinct.
Self-Check — Before You Submit
1. A fictional operator's well team proposes a workover to add capability. The register shows export availability is binding. What should the review do?
- A. Approve it; more capability is always valuable
- B. Ask whether the chain can accommodate it, and whether the workover should wait while export is addressed
- C. Let an optimisation model choose and accept its output
- D. Approve it and report the gross production gain
B. Ask whether the chain can accommodate it, and whether the workover should wait while export is addressed — Additional well capability is valuable only to the extent the remainder of the chain can accommodate it. Approving on capability alone is the tempting answer, and it may simply move the queue.
2. In the book's Friday allocation meeting, the refinery can accept 600 barrels, export 500 and storage 300. Why is it wrong to instruct production of 1,400 barrels?
- A. The capacities may share tanks, pumps and the same transfer window, so they cannot simply be added
- B. The arithmetic is wrong
- C. Storage can never be used for production surplus
- D. The refinery always offers the better contribution
A. The capacities may share tanks, pumps and the same transfer window, so they cannot simply be added — The sum is correct on paper, so the arithmetic is not the problem. The capacities are not independent; the team must first establish what they share.
3. A tool recommends an intervention on an asset where your company holds a non-operated interest. What is the right course?
- A. Instruct the field team directly, since the evidence is strong
- B. Book the expected gain in the company's entitlement immediately
- C. Ignore it, since non-operated assets cannot be influenced
- D. Pursue it through partner governance, technical assurance and approval, recording it as a recommendation, not a commitment
D. Pursue it through partner governance, technical assurance and approval, recording it as a recommendation, not a commitment — An AI recommendation does not change the relationship with the operator. Ignoring it is also wrong: the company exercises its contractual rights through partner governance.
Today’s Assignment
Draw the route from reservoir, or your company's first activity, to payment for your decision, name its binding constraint with evidence, and draft that constraint's entry in a single constraint register.
- The route. From reservoir or first activity to payment: equipment, storage, transport, contractual handoffs and approvals. Not an organisational chart.
- The binding constraint. The one condition limiting what your decision can deliver for the period you examine, its type, the evidence, and what you expect to bind next.
- The register entry. Asset or handoff, observed restriction, evidence, operating consequence, owner by role, proposed resolution, approval required and time basis.
- Interests and quantities. Whether the route runs through operated or non-operated interests, and how gross output, your company's entitlement and cash received differ for this decision. A service company names the counterpart constraint it depends on.
- The honest part. Where your company's departmental measures report success while the system loses value, or a constraint carried forward out of habit. No AI in use yet is normal.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 3 — The route to payment, the binding constraint and our constraint register.
How this is marked — the rubric
| Draws the route through handoffs and approvals to payment, not by department | 25 |
| Names the binding constraint with its type, evidence and what binds next | 30 |
| Completes the register entry with an owner by role and a time basis | 25 |
| Distinguishes operated and non-operated interests and gross, entitlement and cash, and names a departmental success hiding system loss | 20 |
| 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 — The Mandate and the System 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
Route-to-payment mappingFinding the binding constraintRunning a constraint registerIntegrated allocationNon-operated governance
Part One · Leadership and the Operating FoundationSESSION 4
Building an Operating Record Leaders Can Trust
Why This Matters
A chief executive asks which daily production number is correct and receives three. None is necessarily wrong; they describe different things. This session is about the record behind the decision you chose: what each critical number means, who owns its quality, how disagreements are resolved before anyone interprets them, and how the agreement about that record is written down so the same argument does not return every month-end. Trust in a record does not require perfect data. It requires that your team understands what the record represents, where it came from, its limitations and whether it is adequate for the decision. Leaders can approve with imperfect evidence. They cannot defend a choice when the limitation was hidden.
25-Minute Module
0:00 – 6:00
Define the decision before the dataset
- One system reports gross field production, another net entitlement, a third includes condensate, and a field day and an accounting day close at different times. The first correction is definitional: no amount of model training can make unlike measures directly comparable.
- Start with the question. A maintenance intervention and a sales reconciliation may share asset identifiers but need different evidence. A large repository helps only if it preserves the relationships needed to reconstruct a decision.
- For each important signal, define the unit, time basis, location, measurement method and responsible owner. Record whether the value is measured, estimated, allocated or manually entered. Where it has been adjusted, keep the original and the reason.
- A pump named differently in the historian, maintenance system and procurement catalogue is hard to analyse as one asset. A controlled hierarchy should connect equipment to its tags, components, spares and work history.
- THE DOCTRINE
A trustworthy operating record does not mean every data point is perfect. It means the team understands what the record represents, where it came from, its limitations and whether it is adequate for the decision.
Name the decision first; the dataset will be smaller than you feared.
6:00 – 11:00
Measurement quality is an operating responsibility
- Engineering should own the suitability of measurements used for engineering decisions. The data team can reveal anomalies; it cannot establish that a drifting transmitter is fit for its intended purpose.
- Nigeria's Upstream Petroleum Measurement Regulations 2023 include a requirement for an approved metering plan for petroleum production from a producing licence or lease area (NUPRC, Nigerian Upstream Petroleum Measurement Regulations 2023, regulation 3). The book reads this carefully: it places measurement within a formal operating framework; it is not evidence that a virtual estimate can automatically replace an approved measurement arrangement.
- An estimate can help detect a discrepancy or guide a test. Your company must still know which quantities are accepted for allocation, custody transfer, accounting and regulatory reporting.
- Calibration status and uncertainty should travel with critical measurements. When an instrument is unavailable, record the substitute method and its approved period.
A number stripped of its context looks more precise than the evidence allows.
11:00 – 17:00
Reconciliation before interpretation, and the data contract
- An anomaly is a prompt to investigate. It is not a diagnosis. A gap between well estimates and facility measurement could come from instrumentation, allocation, changing fluid properties, time alignment or an actual loss.
- I recommend a defined exception process. The system identifies the discrepancy. A named reviewer checks the evidence. The owner decides whether correction is justified. The corrected record keeps its source and approval. The analytical team then assesses whether the change affects a model or a reported result.
- THE DOCTRINE
Automatic correction can suit tightly defined formatting issues. A system should not quietly replace an inconvenient reading because it differs from the pattern it expects.
- A data contract is, in plain terms, a written agreement about the records a particular use needs: source, definitions, delivery timing, quality checks, ownership, access, and the response when the agreement is not met. For production surveillance it should state how missing values are treated and when the application must withhold a recommendation.
- Keep it realistic: requiring continuous high-quality data from an intermittently connected field without funding instrumentation and communications creates routine non-compliance. Make gaps visible.
- AT THE ASSET · THE WORKED CASE
- Asset. On a fictional onshore field a well meter begins to drift. The platform flags the anomaly; the instrument engineer, as technical owner, decides whether it remains fit for purpose and records the substitute method while it is recalibrated.
IN THE OPERATING REVIEW · THE PARALLEL- Review. At the weekly review the same gap appears as a deferment figure. The review does not accept or reject it on confidence. It asks which quantity is accepted for allocation and reporting, and who approved the correction.
The exception process is where trust is earned, one discrepancy at a time.
17:00 – 22:00
The path from source to decision, and when the daily totals disagree
- A decision should be reconstructable: which records, document versions and assumptions were used. For generative systems a fluent explanation is not a source; an assistant should cite the current approved document and passage.
- The executive measure of data quality is not the number of records stored. Track unresolved critical discrepancies, missing signals, correction time and the proportion of records with a valid source and owner.
The book's invented example Figure What it means Field report 8,200 barrels Gross field production before a later adjustment Accounting report 6,560 barrels An 80% entitlement applied to the gross quantity (8,200 × 0.80 = 6,560) Transfer ticket 7,900 barrels A different time window, including a tank movement - None of the three can be selected as the universal daily production number without reconciliation. Create a record for each quantity with definition, unit, reference conditions, start and end time, source, adjustment and owner. Reconcile inventory movement separately from production, and show any unresolved measurement difference rather than burying it in a balancing entry.
- THE DOCTRINE
A model predicting the company's sales entitlement should not silently train on gross production. Resolve the definitions once and maintain them in the approved data contract.
Which number is correct is the wrong first question; what each number means is the right one.
22:00 – 25:00
Write it down, and Section 4
- Before you close the laptop, write the one record your decision depends on most and three things you are not sure of about it: who owns it, what it means, and how it has been adjusted.
- Section 4 of your plan is The record behind our decision and its data contract.
- Trust develops when the company handles uncertainty openly. A bounded decision with a known limitation is defensible; a decision with a hidden one is not.
Make the gaps visible before anyone builds on them.
Key Terms
Data contract
A clear agreement about the records needed for a particular use: source, definitions, delivery timing, quality checks, ownership, access and the response when the agreement is not met.
Data provenance
The record of information's origin and subsequent handling.
Exception process
The route for a discrepancy: identified by the system, checked by a named reviewer, corrected only on the owner's decision, with source and approval kept.
Approved metering plan
A requirement of the Nigerian Upstream Petroleum Measurement Regulations 2023 (NUPRC, regulation 3) for petroleum production from a producing licence or lease area. A virtual estimate does not automatically replace it.
Self-Check — Before You Submit
1. A chief executive is shown 8,200, 6,560 and 7,900 barrels for the same day and asks which is correct. What is the right first response?
- A. Use the accounting figure, because finance relies on it
- B. Average the three figures to smooth the error
- C. Establish what each number means, including its definition, time window and adjustment, before reconciling
- D. Ask the data team to train a model to predict the true figure
C. Establish what each number means, including its definition, time window and adjustment, before reconciling — The first correction is definitional. The accounting figure is tempting, but it is an entitlement quantity, and no model can make unlike measures comparable.
2. A vendor proposes that its virtual flow estimate replace the approved metering arrangement. How does the book treat this?
- A. An estimate can help detect discrepancies or guide tests, but the regulation is not evidence that it can automatically replace an approved measurement arrangement
- B. The 2023 regulations approve virtual estimates as substitutes for metering
- C. Replacement is acceptable once the estimate matches the meter for a month
- D. Measurement is a data-team matter, so the vendor and data team can decide
A. An estimate can help detect discrepancies or guide tests, but the regulation is not evidence that it can automatically replace an approved measurement arrangement — The regulation places measurement within a formal framework. Matching for a month sounds like evidence, but the book gives no basis for converting an estimate into an approved arrangement.
3. A data platform automatically replaces pressure readings that differ from the pattern it expects. What should the operations leader require?
- A. Keep it; cleaner data improves models
- B. Disable all automatic processing
- C. Have the data team sample replacements quarterly
- D. Limit it to tightly defined formatting issues; route engineering quantities through a named reviewer and owner, keeping the original
D. Limit it to tightly defined formatting issues; route engineering quantities through a named reviewer and owner, keeping the original — Automatic correction is a poor default for consequential engineering quantities, but disabling everything goes too far: formatting fixes are fine. Quarterly sampling still lets readings be quietly replaced.
Today’s Assignment
Define your decision before the dataset, name the measurement and reconciliation weaknesses behind it in your company, and write a practical data contract for its critical record.
- The decision first. Restate your decision and list the evidence it actually needs, not every record your company holds.
- The critical record. For the signals that matter most: unit, time basis, location, method, owner, and whether measured, estimated, allocated or manually entered.
- Weaknesses. Where ownership of measurement quality is unclear, where numbers disagree, and how discrepancies are resolved today.
- The data contract. Source, definitions, delivery timing, quality checks, owner, access, treatment of missing values, when a recommendation must be withheld, and the response when the contract is not met. Make it one your company can actually maintain.
- The honest part. The disagreement about this record that keeps returning, and why it has not been resolved. No AI in use yet is normal.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 4 — The record behind our decision and its data contract.
How this is marked — the rubric
| Starts from the decision and limits the evidence to what it needs | 20 |
| Defines the critical record fully and names who owns measurement quality | 25 |
| Names concrete reconciliation weaknesses and the exception route that should resolve them | 25 |
| Writes a complete, realistic data contract including missing values and when to withhold a recommendation | 30 |
| 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
Decision-first dataMeasurement ownershipReconciliation disciplineWriting a data contractReconstructable decisions
Part One · Leadership and the Operating FoundationSESSION 5
Connecting Field Systems Without Exposing the Plant
Why This Matters
Connecting systems is where an analytical project quietly becomes an operating dependency. A useful application can read what it needs without gaining any ability to change the plant, and that separation should be a deliberate choice, not an accident of convenience. This session covers what your decision needs from field connectivity, how to keep information separate from control, how the system must behave when the connection fails, what to require of a vendor before signing, and how to prove that you can recover. Each is a leadership question. None should be left for you to infer from a diagram drawn for specialists. The session closes Week 1 and opens your first practice exam.
25-Minute Module
0:00 – 6:00
Information versus control, in plain words
- Plain language first. Information is reading the gauge: seeing what a well, pump or compressor is doing and recording it. Control is turning the valve: changing a setting that alters what the equipment does. An application can be very useful while only reading.
- The book's example: an application connects historian data and maintenance history to prepare an engineering review. It needs no permission to change a control-system setting to do that.
- Corporate systems and operational technology, which monitors or controls physical operations, differ. An office application can usually be restarted without changing the physical condition of a facility. A failure or unauthorised change in a control environment may affect equipment, containment or the ability to respond.
- THE DOCTRINE
Begin with the intended use and its consequence, then provide the minimum access required to perform the approved task. Read-only access should be the starting point for many applications.
- If the business later proposes execution capability, treat it as a new approval with a different consequence. The original approval to analyse data is not sufficient, which is the Autonomy Ladder from Session 1 applied to architecture.
Reading the gauge and turning the valve are different permissions; keep them apart on purpose.
6:00 – 11:00
The OT principles, and imperfect connectivity
- The international guidance Principles of Operational Technology Cyber Security identifies six guiding principles: safety, business knowledge, protection of OT data, network separation, secure supply chains and people. It is designed to help organisations assess decisions affecting OT environments (Australian Signals Directorate and international partners, Principles of Operational Technology Cyber Security, 2024).
- For an operator, an integration proposal therefore needs both cyber-security and operating review. A connection can be technically convenient and still introduce an unacceptable dependency or access route.
- Remote operations may have interruptions, limited bandwidth or delayed synchronisation; design for what the asset can sustain. Local and cloud deployment each solve some support problems and create others. Neither location is inherently correct.
- State what happens when data becomes stale: whether the application stops recommending, shows the last observation with its age, and distinguishes an equipment condition from a communications failure. A missing signal should not be silently interpreted as a healthy asset. Local plant control and approved emergency procedures must continue without the analytical service, and that needs testing.
Design for the connection the field has, not the one in the proposal.
11:00 – 16:00
Connecting the records that matter
- The most useful interfaces often sit between existing systems: a condition signal joined to operating status and maintenance history; a delivery forecast joined to tank availability and transport schedules.
- Begin with a small set of controlled connections. For each, document what it reads or writes, its authorised identity, its data owner and its failure behaviour. An enterprise platform may be justified as the scope grows; it should not be treated as a prerequisite for every bounded improvement.
- THE DOCTRINE
The assistant's conversation history should not become the only evidence of a consequential action. The accepted work order belongs in the approved maintenance system.
- AT THE ASSET · THE WORKED CASE
- Asset. At a fictional swamp location an advisory application reads compressor signals from the historian and work history from the maintenance system, and drafts a work order for engineering review. It has no write access to the control system.
IN THE OPERATING REVIEW · THE PARALLEL- Review. At the operating review the accepted work order is the one in the approved maintenance system, and the decision paper sits in the controlled document repository. The review can reconstruct what was seen, who approved it and what was done without opening the assistant's chat history.
A small set of well-documented connections teaches more than a large platform nobody can explain.
16:00 – 21:00
Vendor exit, and proving recoverability
Before signing, require Why it matters Data access and usable export formats A dashboard is of limited value if the underlying evidence cannot be extracted Model and configuration records, audit logs So decisions and changes can be reconstructed Change approval, notification, validation and rollback A remote update can alter behaviour while field hardware is unchanged Full cost: connectivity, monitoring, security, support, training, data correction, eventual replacement Comparing subscription prices alone understates the commitment A plan for changed terms or withdrawn support So the company can still operate - Exercise the recovery plan: restore records from backup, confirm authorised staff can run the fallback, and test behaviour when an interface or vendor service is down, with the people who would respond. A procedure understood only by its developer is not an operating capability.
- The book's fictional field interruption: an advisory application depends on a remote service, and the connection fails shortly before the shift review. Local control continues under approved arrangements, but the dashboard retains its last forecast. The main risk is now a user mistaking an old recommendation for a current one.
- The remedy: show when each input and recommendation was last updated, and prevent a stale recommendation from appearing current. Rehearse the interruption before commissioning, without granting the vendor new permissions to make the demonstration work.
- THE DOCTRINE
A tool that performs well while connected can still be an unsuitable operating dependency if the organisation cannot recover it or continue without it.
If recovery depends on one supplier employee, that belongs in the support decision.
21:00 – 25:00
Write it down, Section 5, and Practice Exam 1
- The architecture review should end with a plain statement of permitted use and residual dependencies, so leadership can judge whether the benefit justifies them.
- Before you close the laptop, write what your decision needs to read, and whether anything in it would ever need to write.
- Section 5 of your plan is Connectivity, separation from control, vendor exit and recoverability. With it, the Week 1 foundation of your plan is complete.
- Practice Exam 1 opens now — fifteen questions, closed book, on Sessions 1 to 5. Sit it before you start Session 6.
Leadership should never have to infer a dependency from a diagram.
Key Terms
Operational technology
Systems associated with monitoring or controlling physical operations; their protection must reflect operating and safety consequences.
Read-only access
An integration that can read data but cannot change settings or records in the source system. The book's starting point for many applications.
Advisory application
A system that provides information or a recommendation for competent human review within its approved scope.
Recoverability
The demonstrated ability to restore records, operate the fallback procedure and continue when an interface or vendor service fails, proved by exercise, not by statement.
Self-Check — Before You Submit
1. During a pilot of an advisory application, the vendor asks for write access to the control system to act faster. What should the leader decide?
- A. Grant it temporarily for the pilot and review later
- B. Grant it if the vendor holds a recognised certification
- C. Treat it as a new approval with a different consequence, needing cyber-security and operating review, and keep the analytical use read-only
- D. Refuse any connection between the application and field systems
C. Treat it as a new approval with a different consequence, needing cyber-security and operating review, and keep the analytical use read-only — The approval to analyse data does not cover execution. Refusing all connection is the overcorrection: read-only access is the book's starting point, not no access.
2. In the book's field interruption, local control continues but the dashboard still shows its last forecast. What is the main risk?
- A. A user mistaking an old recommendation for a current one
- B. The plant losing local control
- C. The vendor losing the operator's data
- D. The model needing retraining after restoration
A. A user mistaking an old recommendation for a current one — The book names stale recommendations as the main risk. Loss of local control is the tempting answer, but in this case local control continues under the approved arrangements.
3. Two vendors are being compared mainly on subscription price. What is missing from the comparison?
- A. The number of features each dashboard offers
- B. Which vendor has the larger client list
- C. Nothing, provided both meet the technical specification
- D. Connectivity, monitoring, security, support, training, data correction, replacement, data export and recovery arrangements
D. Connectivity, monitoring, security, support, training, data correction, replacement, data export and recovery arrangements — Comparing only subscription prices understates the commitment, and exit terms decide whether the company can retrieve its own records. Features and client lists say nothing about either.
Today’s Assignment
State what your decision needs from field connectivity in your company, separate information from control, and name the vendor-exit and recoverability conditions you will require.
- What the decision needs. The signals and records your decision needs, how current they must be, and what connectivity your asset can actually sustain.
- Information and control. Which interfaces read and which would write, with identity and data owner for each. If any write is proposed, say who would approve it as a new decision.
- When the connection fails. What the application must do when data is stale or missing, and how local control and procedures continue without it.
- Vendor exit and recoverability. The conditions you will require: data export, configuration records and audit logs, change notification and rollback, full cost, and continuity if the vendor withdraws. The recovery test you will require before commissioning, and who takes part.
- The honest part. Where your company today depends on a connection, a vendor or one person it could not do without. No AI in use yet is normal; apply this to the systems your decision already relies on.
- Four hundred to nine hundred words.
YOUR 90-DAY PLAN
Section 5 — Connectivity, separation from control, vendor exit and recoverability.
How this is marked — the rubric
| States the decision's data and connectivity needs realistically for the asset | 20 |
| Separates information from control interface by interface, with identity and owner, and treats any write as a new approval | 30 |
| Specifies stale-data behaviour and how operations continue without the service | 15 |
| Names concrete vendor-exit conditions and a recovery test with the people who will run it | 25 |
| The honest paragraph names a real dependency in the student's own company | 10 |
| 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 Operating Foundation 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
Information versus controlDesigning for imperfect connectivityControlled interfacesVendor exit termsProving recoverability