Instructor Guide: Courses 08 to 10

Series · Northbank case · Glossary · References

Read this guide before teaching. Follow the agent instructor method for pacing, assistance, assessment, and continuity. This guide supplies answer guidance, discussion methods, and assessment examples. Present that guidance to learners after their attempt, as teaching or feedback. Accept different defensible solutions that meet the stated criteria.

Use Packets A, B, and C. Keep Packet D for the later courses unless you declare an additional exercise. Preserve each source's identity, revision, and status.

Course 08: Planning and Managing Project Execution

In Course 08, each lesson keeps the classical method, states the cost that method was answering, shows what agents change, and develops the extended method. The classical answers stay correct: the Northbank totals, sequences, and authority conclusions are the same as before. What the course adds is how the learner plans, directs, and examines the work when agents do much of the production.

Session sequence

1. Ask learners to distinguish a target from a forecast.
2. Examine the order-release prerequisite with the group.
3. Draw C02's dependency sequence.
4. Calculate C03's sequence independently.
5. Compare the baseline and forecast costs.
6. Discuss supplier timing and decision authority.
7. Assign the four practice outputs, prepared with agents, with a brief and an examination note for each.
8. Divide the pre-release activities into production, examination, decision, and external time.
9. Examine the agent's 355,000 return.
10. Review the arrangement designs for A05's impact assessment in pairs.
11. Review the sponsor briefings in pairs.
12. Assess and repair the execution packages, including each learner's defense.

You can divide a session around Steps 5 and 9, but preserve one project basis across the sessions. Avoid introducing unrecorded approvals to make the problem easier.

Steps 3 to 5 deliberately come before agent use, and the learner course asks for the same order. A learner who has calculated the sequence and totals once knows what an agent's return should contain and can recognize a return that departs from it.

Teaching Course 08 with an agent

Ask the learner to identify the cost basis and release prerequisite before they calculate. Have them show the calculation and schedule dependencies. You may independently check the arithmetic with a tool, but your correct total does not establish that they understood which amounts or durations belonged in it.

When a result differs, work with the learner to locate the first changed input, duplicated amount, or unsupported dependency. Explain the relevant method if needed, then have the learner repair their workings and the affected briefing. Preserve the supplied figures and avoid invented savings.

Before ending the exercise, ask what the corrected forecast authorizes. This question connects the calculation to the decision and exposes a budget or readiness inference that correct arithmetic alone would miss.

Assessing agent-assisted work

Follow the teaching guide's method for assessing agent-assisted work. For Course 08, look for these points in particular.

In the briefs, check that C03 is given with its statement that each item includes actual expenditure, and that A05 is given with its proposal status.

In the examination notes, look for checks such as the delivery count from the end of Week 6 and contingency remaining at 20,000.

Criterion 6 has two parts: assess the designed arrangement and the learner's own examination record separately. Criterion 7 is the defense. Useful changed premises include: if C03's items excluded actual expenditure, how would the forecast change? If the supplier confirmed a four-week delivery, which date moves and which decisions remain?

Coordinated-agent comparison

Course 08's added worked arrangement is an optional way to prepare the same C01/C03 comparison. It does not change totals, duration or A01 authority. Ask who supplies the inclusive-cost condition to the child agents and who examines the combined return. If both helpers use a mistaken brief, agreement gives no independent support. Accept one-agent work where it serves the learning purpose better.

Look for the learner's direction, consequential child briefs or recoverable references, the returns relied on and the learner's actual examination. Do not grade a longer message archive as stronger evidence. Apply the existing criteria to those contributions.

Nine keyed knowledge-check answers

1. Readiness and schedule. A date places an activity on a calendar, but readiness requires the inputs, resources, competence, access, and authority needed for its stated work. Evidence must establish their required conditions.
2. Resource conflict. Independent activities can require the same unavailable person or resource; the calculated dependency sequence then overstates possible concurrency. Revise capacity, sequence, or scope within authority.
3. Budget and forecast. A01 gives the sponsor authority over budget changes. C03 is expressly a forecast: it assesses a 265,000 outcome but does not authorize that expenditure or amend A02.
4. Double counting. Adding actual expenditure duplicates cost when the forecast already includes it. C03 states that each forecast item includes recorded actual expenditure. An actual-plus-remaining model is valid only with non-overlapping amounts.
5. Equipment release. Northbank needs the reviewed technical purchase package and the required expenditure authorization. Technical work must address the configuration, voltage discrepancy, operating basis, and relevant requirements. Supplier timing is an input to planning, not approval.
6. Overlap. Define the activity boundaries, prerequisites, resource capacity, safety and technical consequences, authorization, and affected examination. Calculate savings only after the proposed overlap has a supported basis.
7. Agents and examination. When examination limits the activity, more agents increase the rate at which outputs reach the examiner without increasing the rate at which the examiner can check them. Unexamined outputs accumulate as verification debt, and the finish does not move. A sound answer identifies the examiner's capacity as the constraint.
8. Agreement between agents. Agreement shows that the two returns are consistent with each other. It does not show that the shared basis, the brief, or the method is correct, because agents that share a model or a brief can share an error. Independence requires a different basis of examination, such as a differently prepared check or a competent person.
9. Practices that lose or keep their reason. Accept any pair argued from the cost each practice answered. Practices that lose much of their reason include estimating information work from production hours, batching changes to save drafting effort, and counting produced documents as progress. A learner may argue that batching decisions keeps a reason, because it saves the decision-maker's attention; credit that argument. Practices that keep their reason include the separation of budget from forecast, the separation of technical review from expenditure approval, readiness conditions, and the authority to decide scope. A weak answer names practices without connecting them to the cost.

Calculation guidance

Check both totals independently:

The difference is approximately 10.4 percent of the baseline:

25,000 ÷ 240,000 × 100 = 10.4167 percent.

This percentage describes forecast cost growth against the baseline, but it does not establish the cause of each increase beyond the supplied figures.

C02's sequence gives end Week 4 release, end Week 8 delivery, Weeks 9–10 installation, Week 11 commissioning, and Week 12 handover.

C03's sequence gives end Week 6 release, end Week 11 delivery, Weeks 12–13 installation, Week 14 commissioning, and Week 15 handover.

The forecast handover is three weeks beyond the target. The sequence assumes ready site works and no overlap, but neither assumption establishes technical purchase readiness.

The three weeks come from the two-week later release and the one-week longer delivery. Agents cannot shorten the delivery. The records do not show how much of the time before release is document production, so they support no claim that agents would move the release.

For the separate cost example:

Actual 90,000 + remaining 175,000 = expected final cost 265,000.

The agent's 355,000 return adds actual 90,000 to C03's inclusive total, so it contains a duplicate 90,000. Learners should explain the error rather than merely give the corrected total. A sound brief sentence states that each C03 item already includes its recorded actual expenditure and that no actual expenditure is to be added to C03's total.

Sample practice reasoning

A useful purchase-release table separates technical package review, expenditure authorization, and supplier delivery confirmation, and it identifies the responsible contributor and evidence for each condition.

The technical review can identify B01's missing selected curve and configuration. It can also identify A04's capacity and protection limits. It must not claim that C03 resolved those matters.

A credible sponsor briefing states the 18 litres per second scope, the 265,000 forecast, and the end-Week-15 finish. It identifies the budget and target decisions without implying approval.

An option can request investigation of an alternative supplier slot or a bounded overlap, but the learner must state what remains unconfirmed. Accept other options with comparable evidence and treatment of consequences.

The briefing should exclude A05's proposed increase from the stated forecast; an alternative that addresses 22 litres per second needs a separate impact assessment and scope decision.

In the duration exercise, a sound answer places the supplier's delivery in external time, the electrical engineer's determinations and the technical review in examination time, the expenditure authorization in decision time, and the drafting of the review and package in production time. It notes that a sponsor budget decision may also come before release, because the forecast exceeds the authorized 240,000. It states that the records give no durations for these parts before release.

A sound arrangement for A05's impact assessment gives agents bounded drafting and comparison assignments, each with A05's proposal status in the brief. The mechanical and electrical engineers examine the hydraulic and electrical parts, and the project engineer examines the interfaces between them. Only the examined assessment reaches the sponsor, who decides. Useful tradeoffs include parallel drafting against a longer examination queue, and one large assignment against harder examination. Accept other designs that allocate examination by consequence and keep the sponsor's decision with the sponsor.

Assessment guidance and common defects

Use the course's seven criteria. Rate each as meets the criterion, needs repair, or has insufficient evidence.

Return arithmetic errors for repair. Return unsupported schedule savings for a prerequisite analysis. Require clear authority when a proposal changes budget, target, or scope.

Common defects include:

• Adding actual expenditure to C03 a second time.
• Dropping contingency without stating a supported change.
• Treating a forecast as a budget approval.
• Counting delivery weeks from the wrong release date.
• Reporting a supplier promise as technical readiness.
• Using document counts as complete evidence of progress.
• Claiming a schedule or cost saving from agent use without evidence for the part of the duration or cost it changes.
• Briefing an agent with A05 or another record without its status.
• Treating agreement between agents as independent review.
• Proposing more agent capacity where examination, a decision, or delivery limits the finish.
• Examination notes that state a conclusion without saying what was checked.
• A defense that restates the package but cannot explain the consequence of a changed premise.

Ask the learner to show the affected equation or dependency, because a rewritten conclusion alone does not establish that the underlying error is repaired.

Course 09: Project Engineering and Technical Integration

In Course 09, each lesson keeps the classical method, states the cost that method was answering, shows what agents change, and develops the extended method. The classical answers stay correct: the basis positions, the interface needs in B03, the B01 discrepancies, and the authority conclusions are the same as before. What the course adds is how the learner arranges and examines agent cross-checks, keeps interface ownership with people, and keeps configuration control in step with agent-speed change. The course has no separate arrangement lesson; Lesson 5 carries the systems view of the integration work.

Session sequence

1. Ask each discipline role to state its needed inputs.
2. Examine the difference between A03 and A04.
3. Build the technical basis table together.
4. Have learners compare B01 with A02, A03, A04, and B03 without an agent.
5. Allocate two B03 interfaces with supplier and consumer roles.
6. Review B01 as a project submission.
7. Trace one construction activity, one operating activity, and one maintenance activity across disciplines.
8. Assign the practice outputs, prepared with agents, with a brief and an examination note for each.
9. Examine the agent's 600 V interface entry.
10. Discuss the agreement exercise: what the consistency check established and what it missed.
11. Conduct a peer review of the proposed interface assignments.
12. Trace the substitution through affected work, including the basis revision each affected output uses.
13. Review the integration arrangement designs in pairs.
14. Assess and repair the coordination package, including each learner's defense.

Require learners to read the consumer's needs before they propose the supplier's return. This reveals missing conditions that a discipline-only discussion can overlook.

Step 4 deliberately comes before agent use, and the learner course asks for the same unaided comparison first. A learner who has compared B01 with the records once knows what an agent's cross-check should return and can recognize a return that adds a resolution or misses a gap.

Teaching Course 09 with an agent

Play a receiving role and ask the learner what contribution would let that role proceed. Then switch to the proposed supplier and ask what inputs remain necessary. Give only the position the case supplies; a simulated exchange does not create the missing engineering data.

Use the learner's actual interface entry to examine meaning, timing, conditions, and responsibility. Ask how the consumer will determine suitability after receipt. If you provide a model entry as teaching, require the learner to develop and explain another entry rather than treat yours as their completed coordination work.

For a proposed substitution, have the learner trace an affected relationship into connected outputs. Keep the course's construction, operation, and maintenance coverage rather than letting one familiar discipline stand for the whole examination.

When you play a receiving role, you are an agent simulating a Northbank participant. Say so if the learner treats your answers as operations' or maintenance's actual position; Lesson 4 makes the same point about simulated reviews.

Assessing agent-assisted work

Follow the teaching guide's method for assessing agent-assisted work. For Course 09, look for these points in particular.

In the briefs, check that B01 is given with its status (submitted for technical review, not accepted), that A03 is given as an archived reference, and that A04's value is given with its stated limits. Check that the briefs tell the agent not to supply values for A02's unresolved inputs and not to propose a technical resolution.

In the examination notes, look for checks such as each cross-check finding traced to its record, no 600 V motor arrangement proposed as a resolution, C03 not treated as a corrected submission, and every proposed owner a person with the relevant responsibility under A01.

In integration evidence, look for whether the learner asked which basis two consistent outputs shared. A learner who reports "the agent found all contributions consistent" as integration evidence has not met the course's central point.

Criterion 6 has two parts: assess the designed arrangement and the learner's own examination record separately. Criterion 7 is the defense. Useful changed premises include: if a corrected submission offered 600 V motors, which interface entries change and which determinations remain with the electrical engineer? If all three discipline drafts had been prepared from one brief that cited A04, what would a consistency check establish? If an agent propagated a proposed substitution into the civil contribution before the change decision, how would the learner find and reverse it?

Ten keyed knowledge-check answers

1. Pump selection. The 18 litres per second value lacks the established suction conditions, pressure, water properties, and operating cases; A02 expressly leaves these unresolved. A selected configuration and competent technical assessment are also needed.
2. Electrical survey. A04 records a checked 600 V, three-phase, 60 Hz connection observation, but it does not establish spare capacity, protection coordination, or equipment compatibility. Those matters need electrical assessment.
3. Interface assignments. Add supplier and consumer owners, required contribution, conditions, point of need, examination, and closure evidence. B03 supplies needs but no assigned owners or required dates.
4. Brochure limits. A product-family description may omit the selected model, ratings, accessories, materials, operating cases, and settings, but a project review needs the identified configuration and relevant performance data.
5. Operating and maintenance review. Participants can expose access, isolation, alarms, restart, removal, and spare-part needs. Early examination lets these needs affect design before procurement or construction constrains the response.
6. Substitution propagation. Assess affected interfaces and authority. Record the actual decision. Update configuration and dependent work. Confirm consumers adopted the change. Repeat affected examination and retain valid unaffected evidence.
7. Consistency and correctness. The check establishes that the contributions agree with each other and with the shared brief and basis. It does not establish that the basis is correct, that the combined behavior works across the operating cases, or that the owners examined the contributions, because an error in the shared basis appears in every contribution and the check passes. A sound answer names a check with a different basis, such as one briefed from the source records, or a discipline engineer's examination of the shared basis.
8. Cheap checks and ownership. Ownership follows responsibility and competence under A01, which keeps technical contributions with the discipline engineers and interface coordination with the project engineer. Agents hold no appointment. The cost of checking set how often a project could look at its interfaces, so cheap checks change that frequency; they produce findings for the owners and decide nothing. A weak answer says the agent "manages" the interface.
9. Configuration control at agent speed. If propagation is fast and control is slow, outputs on different revisions of the basis exist side by side, a proposal can be adopted before anyone decides it, and adoption goes unconfirmed. Issue each decided change as one identified revision, have each output record the basis revision it used, and check adoption against that revision.
10. A complete-looking submission. Completeness shows that the fields are filled, not that the values are sourced or checked; B01's 480 V was copied from A03. An agent comparison adds a full list of disagreements with the records it was given and a provenance trace. It leaves open the suitability of the configuration, the electrical engineer's determinations that A04 leaves open, and whether the operating cases are established. A competent person issues the disposition.

Sample basis and interface reasoning

An acceptable basis table states A02's 18 litres per second requirement and its unresolved conditions, and it preserves A05 as a 22 litres per second proposal.

It cites A04 for the observed connection and its limitations, identifies B01 as an unaccepted submission, and treats A06 as open applicability work.

The learner must not turn the observed 600 V connection into an assumed approved motor arrangement; the actual compatibility solution requires competent assessment.

Possible interface assignments include:

These are teaching proposals. Accept a different allocation when it covers the needed contribution and preserves competence and decision authority.

Required dates can use points of need because the packet supplies no detailed interface calendar. Do not require learners to invent exact dates.

Supplier questions and change example

Useful supplier questions ask for:

• The selected configuration and its revision identity.
• Performance data for the established operating cases.
• Motor and starting data for electrical review.
• Loads, dimensions, and access data for dependent design.
• Applicable maintenance and spare-part information.

A question can expose that the operating cases are not yet established. The learner should name the project's required contribution before they request supplier evidence against it.

For a proposed motor substitution, trace electrical supply compatibility, capacity, protection, controls, dimensions, supporting arrangements, operating information, and affected tests.

Do not require a particular substitute voltage or design: the assignment assesses the impact route, not an invented equipment solution.

Sample reasoning for the agent exercises

In the unaided comparison, a sound list gives the 480 V against A04's 600 V, traces the 480 V to A03 and its planning-estimate source note, and records the absent selected pump curve, declared operating cases, and complete configuration list. It notes that B03's needs cannot be met from B01 and that A02's operating conditions are unresolved, so B01 cannot be checked against them.

In the interface-entry exercise, the entry converts A04's observation of the existing connection into a design decision ("motors to be supplied at 600 V") that only the electrical engineer's determination of compatibility, capacity, and protection could support. It names disciplines instead of people as owners, gives no point of need, and omits the receiving owner's examination. "Closed" is unsupported because the consumer has not examined the contribution and the coupled motor selection is unresolved. Do not require the learner to call the supplier and consumer direction wrong: for supply conditions, electrical to mechanical matches B03's stated need. A sound brief sentence states that A04 is a checked observation that does not establish spare capacity, protection coordination, or equipment compatibility, and that the agent must not state a motor voltage or mark an interface closed.

In the agreement exercise, the consistency check established that the two sections agree with each other and with A03. It missed that A03 is an archived reference with no field survey basis, and that A04 records 600 V. A provenance trace to each value's source and standing, a check briefed from the source records instead of the production brief, or the electrical engineer's examination of the basis would have found the error.

A sound integration arrangement runs agent cross-checks at each issue of a contributing document, against an identified revision of the basis table, and sends findings to the named supplying and receiving owners, with the project engineer as integration owner. It issues decided changes as identified revisions, has agents propagate them, and checks adoption output by output. It gives at least one check a different basis from production, and it keeps an integrated review by people for the shared basis, the coupled motor and electrical decision, and changeover behavior across the operating cases. Useful tradeoffs include detection speed against the owners' finding queue, and a separately briefed check against the cost of a second assignment. Accept other designs that keep ownership with people and distinguish consistency from correctness.

Assessment guidance and common defects

Use the course's seven criteria. Confirm that both the proposed supplier and consumer can explain the interface's required contribution.

A disposition can change a requirement or limit reliance only within the decision owner's authority. It does not establish unexamined behavior.

Common defects include:

• Giving only discipline names as interface definitions.
• Treating receipt as evidence of suitability.
• Omitting the consumer's examination responsibility.
• Using C03 as a corrected technical submission.
• Assuming automatic changeover is already approved.
• Closing integration findings with a meeting label alone.
• Naming an agent as a supplying, receiving, or integration owner.
• Reporting agreement among agent outputs as integration evidence without identifying the basis they shared.
• Briefing an agent with A03 or B01 without its status, or with A04's value without its limits.
• Accepting an agent output that fills A02's unresolved inputs with typical values.
• Propagating a proposed substitution into dependent work before the change decision.
• Letting each agent assignment produce its own configuration list.
• Treating an agent's simulated operator or maintainer review as operations' or maintenance's review.
• Examination notes that state a conclusion without saying what was checked.
• A defense that restates the package but cannot explain the consequence of a changed premise.

Ask learners to review one construction activity, one operating activity, and one maintenance activity. Each review must trace the activity's requirements, interfaces, and examination; a collection of separate discipline documents cannot substitute for that account.

Course 10: Assurance, Delivery, and Handover

In Course 10, each lesson keeps the classical method, states the cost that method was answering, shows what agents change, and develops the extended method. The classical answers stay correct: what C04 supports, what C05 shows, and the sponsor response are the same as before. What the course adds is the separation of an examination's cost from its independence, the scope of agent-prepared evidence, the identity of an agent-assembled package, and Lesson 6 on designing the assurance arrangement.

Session sequence

1. Ask learners what C04 actually observes.
2. Compare that observation with its cover-sheet claim.
3. Have learners write the finding against the cover-sheet claim without an agent.
4. Distinguish verification and validation using one example.
5. Build a coverage table from the project basis, with columns for who performs each examination and its source of independence.
6. Identify commissioning prerequisites, separating document prerequisites from physical and human ones.
7. Examine C05 from the receiving party's perspective.
8. Assign the four practice outputs, prepared with agents, with a brief and an examination note for each.
9. Examine the agent's C04 summary.
10. Redesign the "verified" check.
11. Review the assurance arrangement designs in pairs.
12. Conduct a receiving-operations role exercise.
13. Review the residual-obligation account.
14. Assess and repair the proposed handover package, including each learner's defense.

Keep the exercise at the level of planning and evidence accounts. It supplies no authority or procedure for operating physical equipment.

Steps 1 to 3 deliberately come before agent use, and the learner course asks for the same order. A learner who has written the C04 finding once can recognize an agent account that overstates C04's scope or turns unexercised scenarios into failures.

Teaching Course 10 with an agent

Ask the learner to state the narrow observation that C04 supports before they evaluate its cover-sheet claim. If they conclude that every untested function failed, explain the difference between missing examination and observed failure, then ask them to repair the finding while retaining valid evidence.

In the receiving-role exercise, ask what the proposed package enables the recipient to understand or do, and which obligations remain. Wait for the learner's explanation and inspect the actual handover account. Do not invent completed training, configuration checks, or an operating authorization to move the scenario forward.

A bounded recommendation can meet the learning criteria while actual project acceptance remains unsupported. Assess that reasoning and the required coverage, rather than rewarding an unqualified “complete” status.

If you are an agent instructor, your own check of a learner's agent-prepared account is not independent of it merely because you are a different agent. Say so when it matters, and point the learner to Lesson 6.

Assessing agent-assisted work

Follow the teaching guide's method for assessing agent-assisted work. For Course 10, look for these points in particular.

In the briefs, check that C04 is given with its status, including that it does not establish that earlier design gaps were resolved, that A05 is given as a proposal, and that the brief forbids "passed", "verified", or "complete" without observation, conditions, and limits. Check that any brief touching a check forbids changing expected values or criteria.

In the examination notes, look for checks against C04's body rather than its cover sheet: one normal transfer condition, all four unexercised scenarios listed and not described as failed, the incomplete configuration list, and the 480 V instruction.

In the arrangement account, look for a stated source of independence for each consequential claim. "A second agent checked it" is not a source of independence; a check briefed from the source records, criteria prepared before reading, a physical test within its scope (whose conditions and expected behavior still come from the basis), or a competent reviewer is.

Criterion 6 has two parts: assess the designed arrangement and the learner's own examination record separately. Criterion 7 is the defense. Useful changed premises include: if C04 had recorded the full hydraulic conditions for its one condition, what would it then support, and what still not? If an approved record of the installed configuration existed, which checks could agents then perform, and which remain with people? If the checking agent had used a different model but the same brief and cover sheet, would the check be independent? If the sponsor authorized a qualified handover, which obligations remain, and who owns each?

Ten keyed knowledge-check answers

1. Early examination. Production can hide features or make correction expensive, and some decisions need examination before purchasing or irreversible work. Plan evidence, competence, independence, and repair time with the work.
2. Verification and validation. Verification asks whether specified requirements are satisfied. Validation asks whether the result is fit for intended use under relevant conditions. A method supports only the questions its basis and coverage can answer.
3. C04 observation. It records 18 litres per second in one normal condition with incomplete hydraulic details. It does not establish performance across the required operating cases or unexercised behavior, and it cannot support complete acceptance.
4. Evidence identity. Identify configuration, revision, starting state, conditions, method, instruments where relevant, observer, expected behavior, actual observation, and limitations. Preserve the source records needed to interpret the conclusion.
5. Commissioning and handover. Commissioning examines readiness and behavior under defined conditions; handover transfers the identified result and responsibilities. Receiving information, training, maintenance arrangements, and the relevant decisions remain distinct considerations.
6. Qualified transfer. The actual requirements and authorized decision must permit the qualification. Preserve the unmet obligation, use limitation, receiving owner, resources, closure condition, and authority. Transfer does not prove fulfilment.
7. Independence of an agent review. Accept any three of: a shared model, a shared brief, a shared reference or oracle, and reading the producer's account instead of the actual candidate. Ways to design independence include briefing the checker from the source records, having the examiner state the criteria before reading the output, withholding the producer's conclusion, using a different method such as a deterministic tool or a physical test, and using a different model or a competent person. A weak answer says only "use a different agent".
8. Consequence and detectability. Cheap checks lower the cost of looking, but they do not remove shared errors, and the competent examiner's attention remains scarce. Consequence and detectability show where an error would cause serious harm without surfacing in time, which is where independent human examination buys the most confidence. Required professional, contractual, and project examinations still apply.
9. A complete-looking package. It establishes that documents have been gathered and are available. It does not identify the delivered result: C05 records no complete approved record of the installed configuration. Before the package can identify the result, the approved configuration record must exist, instructions must be checked against it (C04's 480 V instruction), evidence must be stated within its scope, and the receiving people must confirm that they can use it.
10. What agents leave unchanged. The physical tests that C04 did not exercise (standby changeover, power interruption, alarm response, and maintenance isolation), the commissioning prerequisites on the installation, operations' training and the maintenance spare-parts arrangements that C05 records as incomplete, the decisions of the sponsor, technical authority, and operations lead, and A06's open applicability. These are physical processes, people's learning, and authorities under A01, so faster paperwork does not move them.

Sample findings and coverage

A useful finding cites the complete-acceptance claim on C04 Rev 0's cover sheet. It identifies the single observation, the missing hydraulic conditions, the unexercised scenarios, and the incomplete configuration evidence.

Its consequence is that complete acceptance lacks a supporting basis. Its response includes correcting the claim and arranging the missing definition, examination, and decisions.

Do not accept a finding that declares all untested behavior failed: the report supplies absence of examination, not observed failure in each scenario.

A useful coverage table can include flow, changeover, electrical compatibility, installed configuration, operating instructions, training, and maintenance arrangements. Learners need six identified requirements or readiness conditions.

Flow evidence needs the configuration, suction and discharge conditions, water properties, relevant operating case, measurement method, and other conditions that competent assessment needs. Do not demand invented technical values.

Changeover and abnormal responses need their required behavior established before defining expected test results. A05's automatic-changeover proposal has not amended A02.

A06's open applicability remains visible. The teaching packet cannot establish legal conformity for an unidentified jurisdiction.

Sample handover reasoning

The configuration package should reconcile the installed equipment with the defining records and connect the instructions to that configuration. The earlier 480 V reference requires examination and correction.

Operations needs applicable instructions, training, defined operating conditions, and receiving arrangements. Maintenance needs applicable equipment information and confirmed spare-parts arrangements.

The sponsor's request does not establish complete handover. A01 identifies sponsor, technical, and receiving responsibilities. The proposed account must route each decision to the appropriate authority.

Accept a properly bounded proposal for further examination or qualified delivery only when it preserves every unmet condition and required decision. Do not accept unsupported permission for routine operation.

Identify any authorized change to a requirement; such a change does not prove that the original result met the earlier requirement. Qualified delivery must remain within applicable duties.

An agent-assembled package can meet the criteria only if it binds its contents to an identified configuration or states plainly that the approved configuration record does not yet exist, and lists the sections filled with drafts or summaries.

Sample reasoning for the agent exercises

In the summary exercise, "meets A02's 18 litres per second requirement" overstates C04: A02's requirement applies under operating conditions the project has not established, C04 records one normal transfer condition without the full hydraulic conditions, and C04 states that it does not establish that earlier design gaps were resolved. "Minor documentation updates" misdescribes four unexercised scenarios, whose behavior is unestablished, an incomplete configuration list, an instruction that cites the 480 V submission, and C05's incomplete training and spare-parts arrangements. Sound brief sentences require each statement to carry its observation, conditions, and limits, forbid "meets", "passed", or "complete" without them, and require the agent to list every unexercised scenario and open record.

In the redesign exercise, the "verified" check established that the summary agrees with the cover sheet, both of which overstate C04. It was not independent: it shared the model, it read the producer's framing, and its reference was the cover sheet's claim instead of C04's body. A sound redesign briefs the checker from C04's body and A02, has the examiner prepare the coverage table first, withholds the summary or asks the checker to write its own account for comparison, and assigns examination of the consequential claim to a competent reviewer. Accept other redesigns that change what the error depends on.

A sound assurance arrangement gives flow a competent assessment and a defined performance examination under established conditions; changeover a required behavior defined with operations before a defined functional test with operators taking part; instruction applicability an agent check against the approved configuration record, examined by the responsible engineer; and operating readiness the operations lead's own confirmation, with training completed and a readiness demonstration. Useful tradeoffs include checking widely with agents against the examiners' finding queue, and independent review against scarce competent attention. Accept other designs that set strength by consequence and detectability and state each source of independence.

Assessment guidance and common defects

Use the course's seven criteria. Ask the receiving role to explain what they receive, what they can rely on, and what remains their responsibility.

Common defects include:

• Treating one witnessed measurement as full acceptance.
• Inferring earlier design-gap closure from later observed flow.
• Claiming every untested function failed.
• Omitting the installed configuration from the evidence account.
• Equating uploaded documents with operational readiness.
• Transferring an unmet obligation without its consequences or owner.
• Treating a second agent's agreement as independent review.
• Keeping an agent summary's "verified", "passed", or "complete" without its conditions and limits.
• Presenting an agent-assembled handover package as complete when no approved configuration record exists.
• Accepting agent-drafted expected responses for changeover or other unresolved behavior.
• Counting agent-prepared training material as completed training.
• A brief that lets an agent change expected values or criteria to obtain a pass.
• Naming an agent as the owner of a residual obligation.
• Examination notes that state a conclusion without saying what was checked.
• A defense that restates the account but cannot explain the consequence of a changed premise.

Require repair of the relevant claim, coverage, or responsibility. Preserve valid C04 observations with their limited scope, and do not erase the original evidence while you correct its interpretation.

Continue the learning portfolio

Courses 08 to 10 should produce connected packages: schedule readiness must use the technical interface conditions, and assurance must examine the identified configuration and requirements.

Carry remaining matters into Course 11 without declaring them completed. Retain calculations, decisions, observations, and repaired assignments with their relevant revisions, together with the briefs and examination records that show how the agent-assisted work was examined.

Guidance for the post-acceptance transfer practice

Course 10's added scenario is counterfactual and separate from Northbank. In its first report, the stated behavior appears not to meet an accepted requirement; establish the affected conditions before claiming a diagnosis. In the second, conformity can coexist with failure of fitness for the intended use; the question returns to those responsible for the commitments and reliance. In the third, the requested capability is new scope, not retroactive evidence that the earlier commitment was unmet.

A sound response identifies the accepted subject, source of the report, evidence needed, affected reliance and actual decision authority. Notice, restricted use, repair, withdrawal or a new undertaking depend on consequences and applicable authority, not simply on a new test pass. Accept reasoned combinations of the issues and preserve uncertainty where the scenario lacks facts. Do not supply an actual operating procedure or a simulated approval that changes the Northbank record.

Use this as teaching and changed-premise reasoning, not an eighth Course 10 acceptance criterion. Existing keyed answers and assessed assignment remain unchanged.