11. Digital Supply Chains, Automation and AI

Introduction and Context

Logistics has traditionally been managed through transactions, schedules, physical assets and human judgement. Digital technologies have changed the basis on which these activities are coordinated. A supply chain can now be represented through integrated data, monitored through connected devices, executed through automated systems and supported by algorithms that identify patterns or recommend decisions. The significance of this transformation is not simply that warehouses contain robots or that managers use dashboards. Its deeper significance lies in the movement from fragmented, retrospective control towards connected, predictive and increasingly autonomous decision-making.

A digital supply chain uses information technologies to connect suppliers, manufacturers, logistics service providers, carriers, warehouses, customers and regulators. Enterprise resource planning (ERP) systems provide transaction and master-data foundations. Warehouse management systems (WMS) and transport management systems (TMS) coordinate execution. Control towers provide cross-functional visibility and exception management. Internet of Things (IoT) devices generate real-time operational data, while robotics and automation alter the physical flow of goods. Digital twins model supply-chain behaviour, blockchain can support selected forms of shared traceability, and artificial intelligence (AI) assists with forecasting, optimization, classification, language interaction and decision support.

These technologies should not be evaluated independently. Their value depends on architecture, data quality, process design, organizational capability, cybersecurity and governance. A poorly designed digital system can accelerate incorrect decisions, expose sensitive information or create an expensive form of technological dependence. The central managerial question is therefore not “Which technology should be purchased?” but “Which operational problem should be improved, using what information, under what controls, and with what measurable outcome?”

Learning Objectives

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

1. Explain the role of ERP, WMS, TMS, control towers, IoT, robotics, digital twins, blockchain and AI in supply-chain management.

2. Distinguish transaction systems, execution systems, analytical systems and decision-support systems.

3. Evaluate automation and AI opportunities using operational, financial, organizational and risk criteria.

4. Describe applications of demand sensing, optimization and generative AI in logistics.

5. Assess the importance of data quality, interoperability, cybersecurity, privacy, bias and governance.

6. Develop a phased digital supply-chain implementation approach and suitable performance measures.

7. Apply these concepts to managerial and professional decisions in logistics.

From Transaction Systems to Digital Supply-Chain Intelligence

A useful way to understand digital supply chains is to view them as a layered system. At the foundation are physical assets and processes: inventory, vehicles, docks, storage locations, orders and shipments. Above this foundation are systems that record and coordinate events. Analytical tools transform events into information, and AI or optimization tools convert information into recommendations or automated actions.

The first layer is the system of record. ERP platforms record orders, purchasing, production, finance, inventory ownership and master data. They create a common transactional backbone, although they are not always designed for the detailed, high-frequency decisions made in a warehouse or transport operation. The second layer consists of execution systems. A WMS manages warehouse activities such as receiving, put-away, replenishment, picking, packing and dispatch. A TMS supports carrier selection, load planning, routing, tendering, freight settlement and transport visibility.

The third layer is the system of insight. Business-intelligence platforms, data warehouses, event-streaming systems and control towers aggregate data from multiple sources. The fourth layer is the system of decision and action, which includes optimization engines, machine-learning models, robotic control systems and workflow automation. In an advanced architecture, these layers communicate through application programming interfaces (APIs), integration platforms and standardized data models.

Integration does not mean that every system must be replaced by one universal platform. It means that information is available to the right process at the right time, with clear ownership and consistent definitions. For example, “available inventory” may mean physical stock, unrestricted stock, inventory net of allocations or inventory available at a particular location. Unless the definition is governed, a dashboard may appear precise while supporting contradictory decisions.

Core Digital Platforms in Logistics

Enterprise Resource Planning

ERP systems integrate commercial, operational and financial transactions. In supply chains, they commonly support procurement, sales orders, production planning, inventory valuation, supplier management and financial accounting. Their principal strength is enterprise-wide consistency. An ERP can connect a purchase order to a receipt, invoice, payment and inventory record, thereby providing traceability and financial control.

ERP systems are less effective when used alone for highly granular execution. A warehouse may require location-level task sequencing, labour allocation and real-time equipment coordination, while a transport department may require dynamic routing and carrier tendering. Such functions are often supported by specialized WMS and TMS applications integrated with the ERP.

ERP implementation is therefore a process and governance project, not merely an information-technology project. Decisions about product codes, supplier records, units of measure, inventory ownership and approval rights have long-term consequences. Excessive customization may preserve inefficient legacy practices and make upgrades difficult; excessive standardization may ignore legitimate operational requirements.

Warehouse Management Systems

A WMS controls the flow and storage of goods within a warehouse or distribution centre. It receives expected arrivals, directs receiving and put-away, maintains location records, creates picking tasks, manages replenishment and records dispatch. Advanced systems may use wave planning, batch picking, voice instructions, slotting analytics, labour management and integration with conveyors or autonomous mobile robots.

The performance of a WMS depends heavily on disciplined scanning, accurate location master data, reliable item dimensions and consistent operating procedures. If operators bypass scans or if inventory movements are recorded late, the system’s apparent sophistication cannot compensate for inaccurate execution. A WMS should therefore be designed around physical reality, not only around an idealized process map.

Transport Management Systems

A TMS coordinates transport planning and execution. Typical functions include shipment consolidation, carrier selection, route and load planning, appointment scheduling, tendering, tracking, freight audit and performance analysis. A TMS can support trade-offs among cost, service, capacity, emissions and delivery constraints.

Optimization quality is affected by the quality of constraints. If a planning model does not include vehicle capacity, driver regulations, delivery windows, product compatibility or loading restrictions, its mathematically attractive solution may be operationally impossible. Human planners remain important because unusual events, commercial relationships and incomplete information often require judgement.

Control Towers

A supply-chain control tower is an integrated capability for visibility, analytics, collaboration and exception management. It is not simply a dashboard. A dashboard displays information; a control tower links events to thresholds, responsibilities, workflows and possible responses. For example, a delayed inbound shipment may trigger an evaluation of stockout risk, alternative supply, customer priority and transport options.

Control towers generally require four capabilities: multi-source data integration, a common operational picture, analytical or predictive services, and coordinated response. Their design should prevent information overload. A manager needs to know which exception requires intervention, its probable consequence, the recommended action and the confidence of that recommendation.

Connected Operations: IoT, Automation and Digital Twins

The Internet of Things refers to connected physical objects that collect, transmit or receive data. In logistics, examples include vehicle telematics, warehouse sensors, RFID tags, temperature monitors, GPS devices, smart shelves and equipment condition sensors. IoT data can improve location visibility, cold-chain assurance, asset utilization, maintenance and safety.

The value of IoT is determined by the decision it enables. A temperature sensor is useful when a threshold breach leads to a defined response, such as inspection, quarantine, rerouting or customer notification. Installing sensors without ownership of alerts creates data accumulation rather than operational control. Sensor calibration, battery life, network coverage, device identity and data retention must also be managed.

Automation may be physical, digital or cognitive. Physical automation includes conveyors, automated storage and retrieval systems, sortation equipment, robotic picking and autonomous mobile robots. Digital automation includes electronic order processing, automated freight auditing, workflow routing and robotic process automation. Cognitive automation uses analytical models to classify events, predict outcomes or recommend decisions.

The correct evaluation criterion is not whether automation reduces labour in a particular task. It is whether the redesigned process improves total system performance. Automation may increase throughput but reduce flexibility, create bottlenecks at interfaces or require new maintenance skills. A business case should consider capital expenditure, integration, maintenance, utilization, safety, labour redesign, resilience and recovery from failure.

A digital twin is a dynamic digital representation of a physical asset, process or network, linked to operational data and used for monitoring, simulation or prediction. A warehouse twin may represent storage locations, order profiles, equipment and labour. A transport-network twin may test the effect of depot closures, demand changes or alternative routing. A digital twin is more than a static simulation: its usefulness depends on a continuing relationship between the physical system, its data and the analytical model.

Digital twins support scenario analysis. Before changing a warehouse layout, a manager may test congestion, travel distance and throughput under alternative designs. Before changing inventory policies, a supply-chain team may simulate service levels, working capital and transport requirements. However, a twin is only as reliable as its assumptions, data and calibration. It should support managerial judgement rather than create an illusion of certainty.

Blockchain and Shared Traceability

Blockchain is a distributed ledger technology in which transactions are recorded across a network according to agreed validation rules. In supply chains, it has been proposed for provenance, chain-of-custody records, trade documentation, product authentication and multiparty reconciliation.

Its potential advantage is a shared record that reduces dependence on one organization’s database. This can be useful where parties do not fully trust one another and where traceability is important. Yet blockchain does not automatically prove that a physical event occurred. If incorrect information is entered at the point of origin, an immutable record may preserve incorrect information. The technology also raises questions about governance, participant identity, transaction privacy, interoperability and energy or infrastructure requirements.

Blockchain is therefore most appropriate where there is a clearly defined multiparty problem, a need for tamper-evident records and agreement about standards and data ownership. It is rarely justified merely because a supply chain is complex.

Artificial Intelligence in Supply-Chain Decisions

AI in logistics includes machine learning, optimization, computer vision, natural-language processing and generative AI. These approaches are related but not interchangeable. Machine learning identifies patterns from data; optimization searches for good solutions under defined objectives and constraints; computer vision interprets images; natural-language processing handles human language; generative AI produces text, code, summaries or other content from learned patterns.

Demand Sensing and Forecasting

Traditional forecasting often relies on historical demand, seasonality, promotions and statistical models. Demand sensing supplements these inputs with near-real-time signals such as point-of-sale data, online searches, weather, traffic, promotions, cancellations and inventory availability. Its purpose is to detect short-term changes earlier than a periodic forecast cycle.

Demand sensing does not eliminate uncertainty. A sudden sales increase may represent genuine demand, a promotion, a stockout at a competitor or an unusual one-time event. Models must distinguish demand from observed sales, because stockouts can make sales appear lower than underlying demand. Forecast accuracy should be assessed by product, location, horizon and business consequence rather than by a single aggregate measure.

Optimization

Optimization models support decisions such as inventory positioning, vehicle routing, shipment consolidation, workforce allocation, production sequencing and network design. A basic formulation identifies decision variables, an objective function and constraints. For example, a routing model may seek to minimize cost while satisfying capacity, time-window and service constraints.

Optimization is not automatically superior to human judgement. The model may omit qualitative considerations, use uncertain estimates or be too slow for a real-time decision. Robust and stochastic approaches can address uncertainty, while rolling-horizon planning can update decisions as conditions change. Managers should examine not only the recommended solution but also its sensitivity: what happens if demand, travel time or capacity assumptions are wrong?

Generative AI

Generative AI can draft supplier communications, summarize exception reports, translate operating instructions, create training material, assist with data queries and provide natural-language access to logistics information. It can also help analysts write code or formulate initial scenarios. In a control-tower environment, a generative assistant might summarize late shipments, identify affected customers and present approved response options.

Generative AI should not be treated as an autonomous source of truth. It may produce inaccurate statements, omit important context or reveal confidential information through insecure use. Applications should use controlled data sources, access permissions, audit logs, human review and clearly defined boundaries. It is generally safer to use generative AI first for retrieval, summarization and drafting than for unsupervised execution of high-consequence decisions.

Data, Interoperability and Model Governance

Data is the operating material of a digital supply chain. Important categories include master data, transactional data, event data, sensor data, external data and model-output data. Master data includes item, location, supplier, customer, carrier, vehicle and packaging records. Event data records actions such as order creation, picking, loading, departure, arrival and delivery. Each category has different quality, frequency, ownership and retention requirements.

Common data-quality problems include duplicate records, missing values, inconsistent units, incorrect timestamps, stale locations, ambiguous product identifiers and unrecorded exceptions. Data governance assigns ownership and establishes rules for definition, validation, access, correction and retention. A data-quality programme should prioritize fields that affect operational or financial decisions rather than attempting to perfect every data element at once.

Interoperability allows different systems and organizations to exchange meaningful information. APIs, electronic data interchange, event standards and common identifiers can reduce manual re-entry. Integration design should address not only technical connectivity but also semantic consistency. Two systems may successfully exchange a field called “delivery date” while using different meanings—planned date, promised date, estimated date or actual date.

AI governance extends data governance. A model should have a stated purpose, accountable owner, documented inputs, performance measures, validation procedures, monitoring rules and a process for retirement. Model drift occurs when relationships in the operating environment change. A demand model trained during stable conditions may perform poorly after a major market disruption. Monitoring should include accuracy, false positives, bias, data drift, user overrides and business outcomes.

Cybersecurity, Privacy and Responsible Use

Digital logistics increases the attack surface because more systems, devices, suppliers and interfaces are connected. Threats include ransomware, credential theft, manipulation of shipment data, denial of service, unauthorized access to customer information and compromise of operational technology. A disruption to a warehouse-control system or transport platform can have immediate physical consequences.

A sound security architecture uses least-privilege access, multi-factor authentication, network segmentation, encryption, secure configuration, vulnerability management, backups, incident response and supplier assurance. Operational technology should not be connected casually to corporate networks. Business continuity plans must include manual workarounds and recovery priorities for critical logistics processes.

Privacy is relevant when systems process driver locations, employee activity, customer addresses or behavioural information. Organizations should collect only necessary data, define legitimate uses, control access and comply with applicable legal requirements. Transparency is particularly important when employee performance data or automated monitoring is used.

Bias can enter through unrepresentative training data, historical decisions, proxy variables or unequal data quality across regions and suppliers. For example, a carrier-rating model may penalize small carriers because their data is sparse, or a workforce model may allocate undesirable shifts according to patterns reflecting prior inequity. Bias assessment should examine outcomes across relevant groups, while human review and appeal mechanisms should remain available for consequential decisions.

Governance also includes accountability. Responsibility cannot be transferred to a vendor or an algorithm. Managers remain responsible for selecting appropriate use cases, setting approval thresholds and ensuring that employees understand when a recommendation may be challenged.

A Framework for Digital Investment Decisions

A practical evaluation begins with the operational problem. The team should define the current process, its performance gap, affected stakeholders and the cost of inaction. It should then assess whether the problem is caused by inadequate information, poor process design, insufficient capacity, weak compliance or an inappropriate commercial policy. Technology should not be used to conceal a process that has never been standardized.

The next step is to compare solution alternatives. These may include process redesign, better training, improved data discipline, a new application, automation, an analytical model or a combination. Evaluation criteria should include:

• expected service, cost, working-capital and resilience effects;
• implementation effort, integration complexity and total cost of ownership;
• data availability and model reliability;
• scalability, flexibility and vendor dependence;
• cybersecurity, privacy, safety and regulatory exposure;
• workforce impact, skills and change readiness; and
• reversibility if the solution does not perform as expected.

A phased approach is usually preferable. A pilot should have a defined scope, baseline measures, success criteria, operating owner and exit decision. Successful pilots should be industrialized through standard processes, training, support and governance. A technology roadmap should distinguish foundational capabilities—identifiers, connectivity, master data and security—from advanced capabilities such as predictive analytics and autonomous optimization.

Useful measures include order-cycle time, perfect-order rate, inventory accuracy, forecast error, stockout rate, warehouse lines per labour hour, vehicle utilization, cost per shipment, exception-resolution time, system availability, automation utilization and return on invested capital. Measures should be balanced. Maximizing utilization, for instance, may damage responsiveness or resilience if capacity buffers disappear.

Illustrative Case: NorthStar Distribution’s Digital Control Tower

Background

NorthStar Distribution is an illustrative regional wholesaler serving retail and institutional customers from three distribution centres. It uses an ERP system, separate warehouse applications and spreadsheets for transport planning. Customers complain about late deliveries, while managers report that inventory is simultaneously excessive in some locations and unavailable in others.

Situation and Problem

Information is delayed and inconsistent. The ERP shows inventory by facility, but warehouse records are not always synchronized. Carriers send tracking information in different formats. Planners spend much of each morning assembling reports rather than resolving exceptions. Senior management proposes purchasing an AI control-tower platform immediately.

Analysis

A diagnostic review identifies that the principal problems are not initially a lack of sophisticated AI. Product and location identifiers differ across systems; promised delivery dates are not consistently defined; carrier event messages are incomplete; and no team owns exception resolution. Forecasts are also evaluated only at aggregate level, hiding poor performance on high-priority items.

The project team maps the order-to-delivery process, establishes common definitions and creates data ownership rules. It separates three types of exceptions: information exceptions, operational exceptions and policy exceptions. This prevents a late scan from being treated in the same way as a genuine capacity shortage.

Approach and Solution

NorthStar first integrates ERP, WMS and carrier events through APIs and standardized identifiers. It introduces a control-tower workflow that displays shipment risk, inventory exposure and responsible owners. Rules-based alerts are deployed before machine-learning models. A pilot on a selected customer segment uses demand-sensing inputs from point-of-sale data and a routing optimizer that respects delivery windows and vehicle constraints.

A generative-AI assistant is later introduced to summarize daily exceptions using approved data sources. It cannot change orders or release shipments. Every summary links to source events, and planners approve proposed actions. Cybersecurity controls include role-based access, multifactor authentication and separation between analytical systems and warehouse-control systems.

Outcome

The pilot does not produce a dramatic transformation through technology alone. Its main early benefit is that planners spend less time reconciling records and more time resolving high-impact exceptions. Management also discovers that some service failures result from unrealistic customer promise dates rather than transport execution. The project is expanded only after data-quality, workflow and adoption measures meet agreed thresholds.

Learning Points

The case demonstrates that digital maturity is built sequentially. A control tower requires shared definitions, integrated events, accountable workflows and decision rights. AI can improve sensing and prioritization, but it cannot repair undefined processes or unreliable data. Human oversight is particularly important when recommendations affect customer commitments, inventory allocation or safety.

Challenges and Common Mistakes

Organizations frequently underestimate integration and change-management effort. They may purchase applications without redesigning work, select automation without considering volume variability, or launch dashboards with too many alerts. Another mistake is measuring implementation activity—number of devices installed or users trained—rather than operational outcomes.

Vendor lock-in is a strategic risk when proprietary data models or interfaces make migration difficult. A technology portfolio should therefore include contractual provisions for data access, portability, service levels, cybersecurity responsibilities, model transparency and exit support. Resilience also requires avoiding a single digital point of failure.

Automation can create deskilling if employees are excluded from design and no pathway is provided for new roles. Conversely, refusing automation because some jobs will change can preserve unsafe or inefficient work. Responsible implementation engages employees, explains the purpose of monitoring, provides retraining and evaluates effects on workload and safety.

Current Developments and Emerging Directions

Digital supply chains are moving towards event-driven architectures in which systems respond to operational events rather than relying only on periodic batch updates. Edge computing allows selected data processing near vehicles, sensors or warehouse equipment, reducing latency and network dependence. Industry platforms are increasingly combining visibility, planning and execution, although interoperability remains essential.

AI is also moving from isolated prediction models towards decision intelligence: systems that connect prediction, recommended action, simulation and feedback. Multi-agent approaches may coordinate activities across suppliers, warehouses and carriers, but their deployment requires strict boundaries and testing. Sustainability analytics is becoming more integrated with logistics planning, allowing carbon effects to be considered alongside cost and service.

The likely direction is not complete removal of human decision-makers. It is a reallocation of managerial attention. Routine transactions and predictable exceptions may become increasingly automated, while people focus on ambiguous events, stakeholder negotiation, ethical judgement, resilience and system improvement.

Best-Practice Principles

Effective digital supply chains generally follow several principles:

1. Start with a measurable operational or customer problem.

2. Establish process ownership and common data definitions before scaling technology.

3. Design systems around exceptions and decisions, not only around information display.

4. Integrate specialized platforms through governed interfaces rather than creating isolated applications.

5. Pilot in a bounded environment with baseline measures and explicit stop criteria.

6. Keep humans accountable for high-consequence decisions and provide override mechanisms.

7. Treat cybersecurity, privacy, safety and resilience as design requirements.

8. Monitor model performance, bias, drift and business outcomes continuously.

9. Build workforce capability alongside technical capability.

10. Evaluate total cost of ownership, portability and long-term governance before committing to a platform.

Chapter Summary

Digital supply chains combine connected physical operations, integrated transaction systems, analytical platforms and increasingly intelligent decision-support tools. ERP systems provide enterprise records; WMS and TMS coordinate warehouse and transport execution; control towers integrate events and manage exceptions. IoT creates operational visibility, robotics automates physical and digital work, digital twins support simulation, and blockchain may provide shared traceability where governance conditions are suitable.

AI applications include demand sensing, forecasting, optimization, computer vision and generative assistance. Their value depends on appropriate objectives, reliable data, valid constraints, monitoring and human accountability. Digital transformation also creates risks involving cybersecurity, privacy, bias, model drift, vendor dependence and workforce disruption. Successful implementation therefore begins with process clarity and data governance, proceeds through controlled pilots and measures balanced operational outcomes. The strategic objective is not maximum automation but better, more resilient and more responsible supply-chain decisions.

Key Terms

ERP: An enterprise-wide system that integrates core transactional, operational and financial processes.

WMS: A system that manages inventory, locations, tasks and workflows within a warehouse.

TMS: A system that plans, executes, monitors and settles transportation activities.

Control tower: A capability combining integrated visibility, analytics, exception management and coordinated response.

IoT: A network of connected physical devices that collect and exchange data.

Robotic process automation: Software automation that performs structured, repetitive digital tasks.

Digital twin: A dynamic digital representation of a physical asset, process or network linked to operational data.

Blockchain: A distributed ledger that records validated transactions across a participating network.

Demand sensing: The use of near-real-time signals to detect short-term changes in demand.

Optimization: The systematic search for a preferred solution subject to objectives and constraints.

Generative AI: AI that creates text, code, summaries or other outputs from learned patterns.

Master data: Relatively stable reference data describing items, locations, suppliers, customers and other entities.

Interoperability: The ability of different systems to exchange and meaningfully use information.

Model drift: A decline in model performance because operating conditions or relationships have changed.

Data governance: Rules and responsibilities for data definition, quality, access, use and retention.

Digital resilience: The ability of technology-enabled operations to withstand, recover from and adapt to disruption.

Exception management: The prioritization and resolution of events that deviate from planned conditions.

Human-in-the-loop: A design in which people review, approve, override or supervise automated outputs.

Vendor lock-in: Dependence on a supplier that makes switching systems, data or services difficult or costly.

Review Questions

Short Answer Questions

1. What distinguishes an ERP system from a WMS and a TMS?

2. Why is a control tower more than a dashboard?

3. Define a digital twin and state one logistics application.

4. What is demand sensing, and how does it differ from traditional forecasting?

5. Identify two forms of data-quality failure that can affect supply-chain decisions.

Long Answer Questions

1. Explain how ERP, WMS, TMS and control-tower capabilities can be integrated into a digital supply-chain architecture.

2. Evaluate the operational and managerial implications of introducing robotics into a distribution centre.

3. Discuss the uses and limitations of optimization models in transport and inventory decisions.

4. Explain the cybersecurity, privacy, bias and accountability requirements for AI-enabled logistics.

5. Develop a phased implementation framework for a company seeking to introduce AI-supported supply-chain planning.

Analytical / Application Questions

1. A retailer has accurate warehouse inventory records but frequent stockouts in stores. Explain how you would diagnose the problem before selecting a digital solution.

2. A generative-AI assistant recommends rerouting a shipment using incomplete carrier data. What controls should determine whether the recommendation is accepted?

3. A logistics firm must choose between a blockchain traceability project and improved API integration among existing systems. Analyze the criteria you would use to make the decision.

22. Multiple Choice Questions with Answers

Question 1

Which combination best describes the primary role of a supply-chain control tower?

A. Replacing all physical warehouse equipment with robots

B. Combining integrated visibility, analytics, exception management and coordinated response

C. Recording only financial transactions and supplier invoices

D. Generating historical reports without assigning operational responsibility

Correct Answer: B. Combining integrated visibility, analytics, exception management and coordinated response

Question 2

Why may a mathematically optimal transport plan still be unusable?

A. Optimization models can never include constraints

B. The plan may omit real operational constraints such as capacity or delivery windows

C. Human planners are always faster than software

D. Transport costs are unrelated to routing decisions

Correct Answer: B. The plan may omit real operational constraints such as capacity or delivery windows

Question 3

Which situation most clearly illustrates the value of demand sensing?

A. Using only last year’s annual demand to set a five-year network design

B. Detecting a short-term demand change using point-of-sale and promotion data

C. Replacing all inventory records with blockchain entries

D. Auditing historical invoices after a shipment is delivered

Correct Answer: B. Detecting a short-term demand change using point-of-sale and promotion data

Question 4

A company installs temperature sensors but has no process for responding to alerts. What is the main weakness?

A. The sensors automatically reduce product shelf life

B. Data is being collected without a defined operational decision or owner

C. IoT devices cannot be used in cold-chain logistics

D. Temperature information is relevant only to finance departments

Correct Answer: B. Data is being collected without a defined operational decision or owner

Question 5

Which statement best explains why blockchain does not guarantee physical product authenticity?

A. Blockchain records cannot be shared among organizations

B. A blockchain record may preserve incorrect information entered at the source

C. Blockchain cannot record timestamps

D. Blockchain is designed only for internal warehouse transactions

Correct Answer: B. A blockchain record may preserve incorrect information entered at the source

Question 6

Which practice is most appropriate when introducing generative AI into a control-tower environment?

A. Allowing the model to release shipments without review

B. Using approved data sources, access controls, audit logs and human approval for consequential actions

C. Providing the model with unrestricted confidential data

D. Evaluating the system only by the fluency of its text

Correct Answer: B. Using approved data sources, access controls, audit logs and human approval for consequential actions

Question 7

What is the most likely consequence of inconsistent definitions of “available inventory” across systems?

A. All systems will automatically resolve the discrepancy

B. Managers may make apparently precise but contradictory decisions

C. Warehouse picking will become fully autonomous

D. Forecast error will become irrelevant

Correct Answer: B. Managers may make apparently precise but contradictory decisions

Question 8

Which example best represents model drift?

A. A model is documented before deployment

B. A demand model performs worse after market conditions change substantially

C. A planner overrides a recommendation for a valid business reason

D. A sensor transmits data at regular intervals

Correct Answer: B. A demand model performs worse after market conditions change substantially

Question 9

A warehouse automation project increases throughput but creates congestion at packing stations. What does this demonstrate?

A. Automation should be evaluated at total-system level rather than by one task metric

B. Higher throughput always proves project success

C. Packing stations are unrelated to warehouse automation

D. Automation eliminates the need for process redesign

Correct Answer: A. Automation should be evaluated at total-system level rather than by one task metric

Question 10

Which sequence reflects a sound digital transformation approach?

A. Purchase advanced AI, then define the business problem and data requirements

B. Install sensors everywhere, then decide who owns the data

C. Define the operational problem, establish data and process foundations, pilot, measure and scale

D. Replace all existing systems before consulting operational users

Correct Answer: C. Define the operational problem, establish data and process foundations, pilot, measure and scale