Lean Planning, Logistics and Supply Networks

Introduction

Lean manufacturing is often introduced as a shop-floor system, but its performance depends equally on planning, logistics, purchasing, warehousing, suppliers, and customer fulfillment. A factory cannot sustain flow if schedules change constantly, materials arrive late, warehouses conceal shortages, or suppliers optimize their own operations at the expense of the total network. Lean planning therefore extends the search for value and the removal of waste beyond the plant boundary.

This chapter examines how lean principles are applied to production planning, material handling, inventory, warehousing, supplier development, and end-to-end supply networks. It also evaluates an important managerial tension: just-in-time (JIT) reduces inventories and exposes problems, but low buffers may increase vulnerability to disruption. Effective lean supply networks consequently require not only efficiency, but also visibility, capability development, risk management, resilience, and disciplined recovery mechanisms.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Explain the relationship between lean planning, pull systems, takt time, flow, and capacity.
2. Apply basic calculations for takt time, kanban quantities, cycle stock, and safety stock.
3. Design lean approaches to material handling, warehousing, and transportation.
4. Evaluate supplier development and network configuration decisions.
5. Assess the benefits and risks of JIT, localization, dual sourcing, and strategic buffers.
6. Develop disruption-management responses that balance efficiency, resilience, cost, and service.

1. Lean Planning as a Flow-Creation System

Traditional planning often treats production, purchasing, transportation, and warehousing as separate functional problems. Lean planning instead asks whether the entire system can deliver the required product, in the required quantity, at the required time, with minimum interruption and minimum total waste.

The central planning principles are:

• Customer-driven demand: planning begins with actual or carefully forecast customer requirements.
• Leveling: production volume and mix are smoothed where feasible.
• Pull: replenishment is authorized by downstream consumption rather than merely by upstream forecasts.
• Flow: work, information, and materials move with minimal waiting and interruption.
• Standardization: stable methods make abnormal conditions visible.
• Short feedback loops: deviations are detected and corrected quickly.
• System optimization: decisions are evaluated by total network performance rather than local utilization.

Lean planning does not mean eliminating all forecasts. Forecasts remain necessary for capacity, workforce, supplier commitments, transportation, and long-lead components. The distinction is that forecasts may guide resource preparation, while actual consumption triggers near-term replenishment.

Takt time and production rhythm

Takt time expresses the pace required to satisfy demand. It is not the same as machine cycle time. Takt is a customer requirement; cycle time is a process capability.

Available production time must be stated in consistent units and should normally exclude planned breaks, but the organization must define consistently whether setup, maintenance, and planned downtime are included. If a line has 420 available minutes per day and demand is 210 units per day:

A process can support demand only when its effective cycle time is no greater than takt time, subject to variability, changeovers, quality losses, and capacity constraints. If effective cycle time is 2.3 minutes per unit against a two-minute takt, the process has insufficient capacity even if its average output occasionally appears adequate.

For a process requiring several operators, the theoretical minimum number of operators is:

where t_i is the standard work content of task i. Because operators are indivisible and tasks have precedence constraints, the practical staffing requirement is usually at least the ceiling of this ratio and may be higher after line balancing.

Heijunka and production leveling

Heijunka reduces the amplification of demand variation by distributing volume and product mix across a planning period. For example, producing a weekly requirement in large product-specific batches may create peaks in component demand, overtime, changeover disruption, and finished-goods inventory. A leveled sequence spreads production more evenly, although perfect leveling is not always economical or technically possible.

Leveling has limits. It may be inappropriate when demand is highly intermittent, products have incompatible processes, changeovers are extremely long, or customer service requirements justify strategic finished-goods inventory. The managerial objective is not uniformity for its own sake, but a controlled rhythm that reduces total system instability.

Planning hierarchy

A robust lean planning architecture connects different time horizons:

Sales and operations planning (S&OP), or integrated business planning, provides a forum for reconciling demand, supply, capacity, inventory, and financial consequences. Lean planning is weakened when commercial functions promise demand without capacity review or when production is measured only by utilization.

2. Pull Systems and Kanban Design

A pull system authorizes upstream activity in response to downstream use. Kanban is one practical signaling method. A kanban may be a card, container, electronic message, barcode event, or other signal that specifies what should be replenished, where, and in what quantity.

A basic container-kanban quantity can be estimated as:

where:

• n = required number of kanban containers, rounded upward to a whole number;
• D = demand rate in units per time;
• L = replenishment lead time in the same time unit;
• α = protection factor for demand or process variation;
• C = quantity per container.

The formula assumes reasonably stable demand, a defined replenishment lead time, consistent container quantity, and a protection factor that is deliberately selected rather than used to conceal poor process control. It is a planning approximation, not a substitute for capacity and risk analysis.

Worked example. A line consumes 480 components per day. Replenishment takes 0.5 day. Management allows a 10% protection factor, and each container holds 40 components.

The system requires 7 containers after rounding upward. The associated maximum authorized inventory is approximately 280 components, assuming all containers are in the loop and no additional stock exists outside the design.

Kanban quantity should be reviewed when demand, lead time, container size, quality performance, or transportation frequency changes. Increasing kanban to compensate permanently for unreliable suppliers may protect the line temporarily but weakens the improvement process.

3. Inventory: Waste, Protection, and Strategic Choice

Inventory performs legitimate functions: it decouples operations, covers lead time, enables economic transportation, supports product availability, protects against uncertainty, and may hedge against supply or price risks. However, excess inventory also creates storage cost, obsolescence, damage, hidden defects, longer response times, and reduced urgency to solve root causes.

Lean inventory management distinguishes several components:

• Cycle stock: inventory required by replenishment lot size and demand between replenishments.
• Pipeline stock: inventory currently in transportation or processing.
• Safety stock: deliberately held protection against uncertainty.
• Anticipation stock: inventory accumulated for known seasonal or planned events.
• Decoupling stock: inventory separating processes with different rates or reliability.
• Obsolete or dead stock: inventory with no economically useful demand.

For a stable item with constant demand and replenishment lead time, average cycle stock under a simple lot-size policy is approximately:

where Q is the replenishment quantity. This assumes a steady depletion pattern, instantaneous replenishment relative to the cycle, and no safety stock. If replenishment is gradual, the average differs and must reflect the production and consumption rates.

A common safety-stock approximation for independent demand during a fixed lead time is:

where SS is safety stock in units, z is the chosen service-factor value, σ_d is the standard deviation of demand per period, and L is lead time in periods. This model assumes demand observations are sufficiently stable, lead time is fixed, demand periods are independent, and the selected distributional approximation is reasonable. If lead time is variable, both demand and lead-time uncertainty should be modeled; the formula above should not be applied without adjustment.

The reorder point under this fixed-lead-time model is:

where d is average demand per period. The first term covers expected lead-time demand; the second protects against uncertainty. Service level must be defined carefully. A cycle-service target concerns the probability of not stocking out during a replenishment cycle, while a fill rate concerns the proportion of demand supplied immediately. They are not interchangeable.

Lean practice does not equate to zero inventory. It seeks the lowest total inventory consistent with safety, quality, service, legal requirements, and economically rational risk protection.

4. Material Handling and Internal Logistics

Material handling is part of the production process, not merely a support activity. Poor handling creates motion, waiting, damage, searching, excess transport, and line stoppages. Lean internal logistics normally uses point-of-use storage, standardized containers, visual addresses, fixed routes, supermarkets, milk runs, and clear replenishment signals.

A supermarket is a controlled inventory location from which downstream processes withdraw material. The withdrawal signal triggers replenishment from an upstream process or supplier. A supermarket is not an unmanaged warehouse: it requires defined locations, maximum and minimum quantities, ownership, cycle counting, and abnormality escalation.

A milk run is a planned route that collects or delivers materials at multiple points on a regular schedule. Compared with independent emergency trips, it can improve vehicle utilization and delivery rhythm. Its design must consider route time, vehicle capacity, loading constraints, traffic, safety, and the risk that a missed route affects several operations.

Internal logistics should be designed around the following questions:

1. What is the consumption point and its required presentation quantity?
2. What signal authorizes replenishment?
3. What route and frequency are appropriate?
4. Where are standards, abnormalities, and shortages made visible?
5. How are damaged, suspect, or obsolete materials segregated?
6. What is the recovery process when the normal route fails?

Line-side inventory must be ergonomically safe. A low-inventory system that increases lifting injuries, collision risk, or repetitive strain is not lean. Safety, quality, and regulatory requirements are constraints on system design, not optional additions.

5. Lean Warehousing

Lean warehousing focuses on reliable flow and accurate information. Core practices include:

• disciplined receiving and verification;
• standardized put-away and location control;
• ABC or risk-based slotting;
• barcode or radio-frequency identification where justified;
• first-in, first-out (FIFO) or first-expiry, first-out (FEFO) control;
• cycle counting based on risk and transaction history;
• cross-docking when it genuinely reduces handling and dwell time;
• visual management of shortages, blocked stock, and quality holds;
• standard work for picking, packing, and dispatch.

ABC classification is useful but incomplete. High-value items deserve financial control, but low-value items can be operationally critical. A better classification may combine annual usage value, lead time, substitutability, safety impact, quality sensitivity, and disruption exposure.

Warehouse performance should be measured as a balanced set rather than by utilization alone. Relevant indicators include inventory accuracy, order-pick accuracy, dock-to-stock time, order cycle time, on-time dispatch, damage rate, space utilization, labor productivity, and safety incidents. Maximizing space utilization may increase travel distance and reduce responsiveness; maximizing labor utilization may create queues and late orders.

6. Supplier Development and Lean Purchasing

Lean purchasing is more than demanding lower prices. It seeks reliable total cost, stable quality, short and dependable lead times, collaborative problem solving, and responsible long-term capability. Supplier development can include:

• shared process mapping and value-stream analysis;
• supplier quality-at-source programs;
• joint root-cause analysis;
• standardized packaging and delivery frequency;
• technical assistance and training;
• common performance definitions;
• early supplier involvement in product and process design;
• development of contingency and recovery plans.

Supplier evaluation should distinguish outcomes from capability. A supplier may currently meet delivery requirements because it carries excessive inventory or uses costly expediting. Conversely, a supplier with strong process capability may temporarily underperform because of a one-time disruption. Audits should therefore examine process control, maintenance, workforce competence, quality systems, capacity flexibility, cybersecurity, financial health, and business continuity.

A total-cost perspective includes purchase price, transportation, duties, inspection, inventory, quality failure, expediting, downtime, administration, environmental impact, and end-of-life obligations. A geographically distant supplier with a low unit price may have higher total cost and greater disruption exposure.

Supplier relationships must also respect competition law, confidentiality, human rights, and anti-corruption requirements. Collaboration does not justify exchanging competitively sensitive information among independent suppliers or imposing unrealistic delivery requirements that transfer hidden costs and risk to vulnerable partners.

7. End-to-End Supply Networks

A supply chain is better understood as a network because firms may have multiple suppliers, plants, distribution centers, logistics providers, customers, and reverse flows. Network design decisions involve location, capacity, transportation mode, sourcing, postponement, inventory positioning, and information architecture.

Figure: Material moves toward customers while consumption information and control signals move upstream through a coordinated network.

Network-wide lean performance requires visibility beyond the immediate supplier. Tier-2 and tier-3 dependencies may contain the true bottleneck, especially for specialized materials, semiconductors, chemicals, tooling, or regulated inputs. Mapping should identify single points of failure, common geographic exposure, shared logistics providers, capacity bottlenecks, and substitutability.

JIT and resilience trade-offs

JIT reduces inventory, shortens lead times, exposes defects, and encourages frequent replenishment. Its vulnerability arises when disruption duration exceeds the system’s recovery capacity. Resilience is the ability to absorb, adapt, and recover while maintaining critical outcomes.

The relevant decision is not “lean or resilient.” A well-designed lean system removes process waste while deliberately protecting high-consequence risks. Possible resilience mechanisms include:

• qualified alternative suppliers;
• regional or multi-regional sourcing;
• strategic safety stock for critical components;
• postponement and modular product design;
• flexible production capacity;
• substitute materials and approved alternate specifications;
• supplier business-continuity plans;
• logistics route alternatives;
• rapid escalation and transparent risk information.

These mechanisms have costs. Dual sourcing may reduce volume leverage and require duplicate qualification. Localization may increase unit cost while reducing transport and geopolitical exposure. Strategic inventory ties up capital and may become obsolete. The correct design depends on disruption probability, impact, recovery time, margin, customer criticality, and substitution options.

A useful managerial classification is:

8. Disruption Management and Control-Tower Thinking

Disruption management should begin before the event. A practical cycle is:

1. Identify: map dependencies, failure modes, and critical flows.
2. Assess: estimate likelihood, impact, detectability, and recovery time.
3. Prepare: qualify alternatives, define inventory policies, and conduct scenario exercises.
4. Detect: monitor leading indicators such as supplier capacity, quality drift, logistics delays, and geopolitical alerts.
5. Respond: protect people, customers, quality, and critical operations in that order.
6. Recover: restore stable flow without normalizing emergency workarounds.
7. Learn: update standards, contracts, designs, and risk registers.

A control tower is a coordinated information and decision capability, not merely a dashboard. It should establish data ownership, exception thresholds, escalation rights, scenario analysis, and action tracking. Technology can improve visibility, but inaccurate master data or ungoverned alerts can create false confidence.

Useful indicators include supplier on-time-in-full performance, lead-time variability, premium freight, inventory record accuracy, forecast error, schedule adherence, first-pass yield, customer fill rate, recovery time, and concentration exposure. Metrics should be segmented by product family, supplier, region, and criticality; network averages can conceal severe local risk.

9. Technology, Analytics, and AI

Digital tools support lean logistics when they make actual conditions more visible and decisions more timely. Examples include warehouse-management systems, transport-management systems, electronic kanban, IoT sensors, digital twins, RFID, advanced planning systems, and supplier portals.

AI and machine learning may assist with demand sensing, ETA prediction, anomaly detection, inventory classification, and disruption scenario generation. However, predictive accuracy depends on data quality, process stability, and changing relationships. Models can reproduce historical bias, overreact to unusual signals, or recommend actions that violate capacity, safety, contractual, or ethical constraints.

Good governance requires human accountability, explainable decision logic for high-impact decisions, access control, cybersecurity, data protection, model monitoring, and clear override procedures. Automation should not conceal abnormalities; it should accelerate their detection and structured resolution.

10. Performance Measurement and Economic Evaluation

Lean supply networks require a balanced scorecard. Cost alone encourages local optimization; service alone may produce excessive inventory. A useful performance architecture combines:

• customer: service level, fill rate, lead time, complaint rate;
• flow: throughput, queue time, schedule adherence, replenishment reliability;
• inventory: turns, days of supply, accuracy, obsolescence;
• quality: defects, returns, first-pass yield, supplier escapes;
• resilience: time to detect, time to recover, alternate-source coverage;
• people and responsibility: safety, training, labor conditions, emissions, compliance;
• financial: total landed cost, working capital, margin, and cash conversion.

Inventory turns are commonly calculated as:

The numerator and denominator must use the same valuation basis. High turns are not automatically superior if they produce stockouts, premium freight, or lost customers.

Working capital is affected by inventory, receivables, and payables. Aggressively extending supplier payment terms may improve the buyer’s short-term cash position while damaging supplier liquidity and increasing network risk. Lean governance therefore evaluates financial performance across the relationship, not only at the focal firm.

11. Illustrative Case: Resilient Lean Supply for Northstar Medical Devices

Background

Northstar Medical Devices, a fictional manufacturer, produces diagnostic modules for hospitals. Its final assembly plant used electronic kanban and daily supplier deliveries. Inventory was low, warehouse space was limited, and service performance had historically been strong.

Problem

A fire at a specialized resin supplier interrupted production. The supplier was the sole approved source, qualification of alternatives required six months, and Northstar held only four days of resin. Expedite shipments did not solve the problem because the material was unavailable. The plant stopped after five days, causing hospital backorders and regulatory escalation.

Analysis

Northstar initially concluded that JIT had failed. A deeper analysis showed several weaknesses: the resin was classified by purchase price rather than criticality; tier-2 dependency was not visible; the business-continuity plan assumed transportation disruption rather than material loss; product specifications allowed no prequalified substitute; and the kanban system had been optimized for normal lead time without a criticality-based protection policy.

Approach

Northstar created a cross-functional risk team involving engineering, quality, procurement, operations, finance, and key customers. It mapped the network to tier 3, identified single-source materials, and classified components using consequence, recovery time, substitutability, and demand. It then:

• qualified a second resin source in a different region;
• redesigned the module to permit an approved alternative resin;
• established a controlled strategic buffer for the critical material;
• required key suppliers to report capacity and sub-tier concentration;
• introduced quarterly disruption simulations;
• created escalation rules for abnormal lead-time and capacity signals;
• retained pull replenishment for routine, interchangeable materials.

The strategic buffer was not applied indiscriminately to every component. For the resin, the target was based on qualification lead time and recovery scenarios rather than an arbitrary number of days. Quality and regulatory approval governed any alternate material.

Outcome

Northstar accepted a modest increase in inventory and qualification cost but reduced sole-source exposure and improved recovery capability. The company also discovered that several apparently lean practices—small deliveries, low supplier margins, and frequent engineering changes—were transferring instability into the network. After redesign, routine materials remained on low-inventory pull systems, while critical materials received differentiated controls.

Learning points

1. Low inventory is not the same as low total risk.
2. Criticality must include consequence and recovery time, not only spend.
3. Resilience is strengthened by design, qualification, visibility, and capability—not inventory alone.
4. JIT is most robust when processes are reliable and exceptions are deliberately governed.
5. Supplier development should address sub-tier dependencies and product design.

12. Implementation Roadmap and Limitations

A practical implementation sequence is:

1. Define customer value, service requirements, and critical products.
2. Map physical, information, financial, and reverse flows.
3. Establish demand, takt, capacity, lead-time, and inventory facts.
4. Stabilize standards, quality, master data, and supplier interfaces.
5. Introduce pull and replenishment controls in a bounded pilot.
6. Improve material presentation, routes, packaging, and warehouse accuracy.
7. Develop suppliers through joint problem solving and capability reviews.
8. Segment inventory and sourcing policies by risk and criticality.
9. Build disruption scenarios, alternate options, and escalation routines.
10. Audit outcomes and revise policies through learning cycles.

Lean transformations commonly fail when leaders implement kanban before stabilizing quality, use inventory reduction as the primary target, ignore demand volatility, or punish people for revealing problems. They also fail when local lean projects optimize a plant while increasing supplier overtime, transport emissions, or customer shortages.

Environmental and social effects require explicit consideration. Frequent small-lot deliveries may improve flow but increase vehicle emissions. Outsourcing may lower purchase price while weakening labor standards or increasing supply-chain opacity. Reusable packaging can reduce waste but may create reverse-logistics burdens. Life-cycle analysis, worker consultation, legal compliance, and responsible sourcing should accompany operational improvement.

Chapter Summary

Lean planning extends lean thinking from the production line to the entire supply network. Takt time links available capacity to customer demand, while leveling reduces instability in volume and mix. Pull systems and kanban connect replenishment to consumption, but their quantities must reflect demand, lead time, container size, and controlled protection against variation.

Inventory is neither inherently waste nor inherently protection. Cycle stock, pipeline stock, safety stock, anticipation stock, and decoupling stock serve different purposes and should be governed differently. Lean logistics uses point-of-use delivery, supermarkets, milk runs, standard containers, accurate warehouse systems, and visual abnormality management. Supplier development emphasizes process capability, quality at source, reliable delivery, total cost, and ethical collaboration.

JIT can expose problems and reduce waste, but resilience requires deliberate protection against high-consequence disruption. Alternative suppliers, localization, strategic buffers, postponement, flexible capacity, and network visibility involve trade-offs. Effective supply-network management therefore combines flow efficiency with risk segmentation, digital visibility, disciplined governance, and continuous learning.

Key Terms

Takt time: The pace of production required to meet customer demand.

Cycle time: The time required for a process or workstation to complete one unit.

Heijunka: Production leveling by volume and, where practical, product mix.

Pull system: A system in which downstream consumption authorizes upstream replenishment.

Kanban: A physical or electronic signal authorizing production or movement.

Supermarket: A controlled inventory location used to support pull replenishment.

Milk run: A scheduled route serving multiple pickup or delivery points.

Cycle stock: Inventory used during normal replenishment cycles.

Safety stock: Inventory held to protect against uncertainty.

Pipeline stock: Inventory currently moving or being processed.

Reorder point: Inventory position at which replenishment is initiated.

Service level: A specified measure of fulfillment or stockout protection.

Supplier development: Activities that improve a supplier’s capability and performance.

Total cost of ownership: The complete cost of acquiring, using, supporting, and disposing of an item or service.

Supply-network visibility: The ability to identify and understand flows, dependencies, and conditions across network tiers.

Resilience: The capacity to absorb, adapt to, and recover from disruption.

Postponement: Delaying final differentiation or configuration until demand is clearer.

Control tower: A coordinated capability for visibility, exception management, and supply-network decisions.

Review Questions

Short Answer Questions

1. Distinguish takt time from cycle time.
2. What is the purpose of a supermarket in a pull system?
3. Identify three types of inventory and explain their purposes.
4. Why can a low purchase price fail to represent low total cost?
5. State two limitations of applying a basic safety-stock formula.

Long Answer Questions

1. Explain how heijunka, pull replenishment, and capacity planning interact in a lean production system.
2. Discuss how lean warehousing can improve flow without sacrificing accuracy, safety, or service.
3. Evaluate the role of supplier development in extending lean principles beyond the focal factory.
4. Compare the efficiency benefits and resilience risks of JIT inventory policies.
5. Explain how a firm should design a balanced performance-measurement system for an end-to-end supply network.

Analytical / Application Questions

1. A line has 360 available minutes per day and demand of 180 units per day. Calculate takt time. If the effective process cycle time is 2.2 minutes per unit, assess whether the process can meet demand and explain your conclusion.
2. A component has daily demand of 300 units, a replenishment lead time of 0.5 day, a protection factor of 15%, and containers holding 25 units. Calculate the required kanban quantity using the formula in this chapter and interpret the result.
3. A company is considering replacing a sole domestic supplier with a lower-cost overseas supplier. Develop a decision framework that evaluates purchase price, lead time, inventory, quality, transportation, carbon impact, disruption exposure, qualification time, and customer consequences.

Multiple Choice Questions with Answers

Question 1

A line has 420 available minutes per day and customer demand is 210 units per day. What is the takt time?

A. 0.5 minute per unit

B. 2 minutes per unit

C. 210 minutes per unit

D. 630 minutes per unit

Correct Answer: B. 2 minutes per unit

Question 2

Which statement best describes a pull system?

A. Production is authorized solely by the annual forecast

B. Upstream production responds to downstream consumption

C. Every process maximizes its local utilization

D. Inventory is eliminated regardless of risk

Correct Answer: B. Upstream production responds to downstream consumption

Question 3

Demand is 480 units per day, lead time is 0.5 day, protection factor is 10%, and container size is 40 units. How many containers are required using the chapter formula?

A. 5

B. 6

C. 7

D. 12

Correct Answer: C. 7

Question 4

Which inventory type protects against uncertainty in demand or replenishment?

A. Safety stock

B. Obsolete stock

C. Cycle stock only

D. Finished packaging waste

Correct Answer: A. Safety stock

Question 5

What is the principal purpose of a milk run?

A. To eliminate all transportation activity

B. To create a planned route serving multiple material points

C. To increase emergency shipments

D. To replace supplier quality management

Correct Answer: B. To create a planned route serving multiple material points

Question 6

Why should supplier classification include criticality rather than purchase price alone?

A. Low-value items can never disrupt production

B. Criticality reflects consequence, substitutability, and recovery difficulty

C. Purchase price is unrelated to financial control

D. Critical suppliers always have the highest prices

Correct Answer: B. Criticality reflects consequence, substitutability, and recovery difficulty

Question 7

Which action most directly improves resilience for a sole-source critical component?

A. Removing all quality qualification requirements

B. Increasing product variety without analysis

C. Qualifying an alternate source or approved substitute

D. Measuring only supplier unit price

Correct Answer: C. Qualifying an alternate source or approved substitute

Question 8

What is a key limitation of the fixed-lead-time safety-stock formula?

A. It cannot use units

B. It assumes fixed lead time and stable, independent demand conditions

C. It always produces zero safety stock

D. It applies only to finished goods

Correct Answer: B. It assumes fixed lead time and stable, independent demand conditions

Question 9

Which warehouse measure is most likely to encourage harmful local optimization if used alone?

A. Inventory accuracy

B. Order-pick accuracy

C. Space utilization

D. Safety incidents

Correct Answer: C. Space utilization

Question 10

Which statement about AI in lean supply networks is most accurate?

A. AI removes the need for data governance

B. AI guarantees correct disruption decisions

C. AI can support prediction and anomaly detection but requires oversight and quality data

D. AI should conceal abnormalities from operators

Correct Answer: C. AI can support prediction and anomaly detection but requires oversight and quality data