Chapter 36 Exercises: Industry Applications of AI

Chapter 36 Exercises: Industry Applications of AI

Section A: Cross-Industry Pattern Recognition (Exercises 1-5)

Exercise 1: Capability Mapping

For each of the following AI applications, identify the primary universal AI capability (prediction, optimization, NLP, computer vision, recommendation, anomaly detection) it relies on. Then identify a parallel application in a completely different industry that uses the same capability.

(a) A hospital uses an ML model to predict which patients admitted through the emergency department are at highest risk of readmission within 30 days.

(b) A wind farm operator uses sensor data to identify turbines that are likely to experience gearbox failure within the next 60 days.

(c) A law firm uses an NLP system to classify 2 million documents as responsive or non-responsive during litigation discovery.

(d) A food manufacturer uses cameras on its packaging line to detect misaligned labels, damaged seals, and foreign objects.

(e) An adaptive learning platform adjusts the difficulty and sequencing of math problems based on each student's demonstrated mastery.

(f) A central bank uses ML to identify unusual patterns in interbank wire transfers that may indicate money laundering.

Exercise 2: The Maturity Framework

Using the industry AI maturity framework from this chapter (Leaders, Fast Followers, Emerging Adopters, Early Stage), complete the following:

(a) For each maturity tier, identify one additional industry not discussed in this chapter and justify your classification based on data readiness, organizational readiness, and regulatory environment.

(b) Select an industry currently classified as "Early Stage." Describe three specific conditions that would need to change for that industry to advance to "Emerging Adopter" status within five years.

(c) Tom argues that financial services leads AI maturity primarily because of competitive pressure, not because of data readiness. NK argues that data readiness is the primary driver. Who is more correct? Build a one-paragraph argument for each position, then state your own view with justification.

(d) Lena observes that heavily regulated industries have better AI governance but slower adoption. Is this a necessary tradeoff, or can regulation accelerate AI adoption under certain conditions? Provide an example from the chapter to support your argument.

Exercise 3: Transfer Analysis

You are an AI consultant who has spent five years building churn prediction models in the telecommunications industry. You have been hired by a hospital system to reduce patient attrition — patients who leave the health system for a competitor or simply stop seeking care.

(a) Identify five elements of your churn prediction expertise that transfer directly to the healthcare context. Be specific about which techniques, frameworks, and approaches apply.

(b) Identify five elements that do not transfer — aspects of healthcare that require entirely new knowledge, data sources, or approaches.

(c) What regulatory constraints apply to healthcare patient data that do not apply to telecom customer data? How do these constraints affect your model design?

(d) A hospital administrator says, "We want you to predict which patients will leave and then offer them incentives to stay — just like telecom companies offer retention deals." Explain why this framing is problematic in a healthcare context and propose an alternative framing that achieves the same goal (retaining patients) without the ethical concerns.

Exercise 4: The Six Capabilities Deep Dive

Select two of the six universal AI capabilities (prediction, optimization, NLP, computer vision, recommendation, anomaly detection). For each capability:

(a) Describe the underlying technical approach in two to three sentences, referencing the relevant earlier chapter in this textbook.

(b) Identify three industries from this chapter where the capability is deployed and describe the specific use case in each.

(c) For one of the three use cases, identify the most significant ethical risk and propose a mitigation strategy that draws on concepts from Part 5 (Chapters 25-30).

(d) Assess whether the capability is more mature (closer to commodity) or more emerging (still requiring custom development) across the industries you identified. What factors determine the maturity level?

Exercise 5: Industry Analogy Framework

Professor Okonkwo says, "The best AI strategists think in analogies — they see how a solution in one industry can be adapted to another." For each of the following pairs, describe how an AI application from Industry A could be adapted for Industry B. Identify what would need to change in terms of data, features, evaluation metrics, regulatory compliance, and deployment context.

(a) Industry A: Retail dynamic pricing. Industry B: Hospital resource allocation.

(b) Industry A: Manufacturing quality inspection (CV). Industry B: Agricultural crop disease detection (CV).

(c) Industry A: Financial services fraud detection. Industry B: Academic integrity violation detection.

(d) Industry A: Media content recommendation. Industry B: Adaptive learning content sequencing.

Section B: Industry Deep Dives (Exercises 6-10)

Exercise 6: Financial Services AI Audit

A mid-size regional bank (30 branches, $5 billion in assets) is considering its first significant AI investment. The bank's CEO has read about AI in financial services and wants to "catch up" with larger competitors.

(a) Propose a prioritized list of three AI use cases for the bank, ranked by expected ROI, implementation feasibility, and risk. For each, identify the data requirements, the estimated implementation timeline, and the regulatory considerations.

(b) The bank's compliance officer is concerned about the explainability of ML-based credit scoring models. Draft a one-paragraph response that addresses her concerns, referencing the explainability frameworks from Chapter 26.

(c) The bank currently uses a rule-based system for AML compliance that generates a 95% false positive rate. A vendor offers an ML-based alternative that promises to reduce false positives to 40% while maintaining detection accuracy. What questions should the bank ask the vendor before making a purchasing decision? List at least six.

(d) The bank's board wants to understand how AI compares to hiring additional human analysts. Using the ROI framework from Chapter 34, outline the factors that should be included in a total cost of ownership comparison between AI and human alternatives for fraud detection.

Exercise 7: Healthcare AI Opportunity Assessment

You are the newly appointed Chief Digital Officer at a 500-bed academic medical center. The hospital has an electronic health record system (Epic), a PACS imaging system, and a research division with data science capabilities.

(a) Using the framework from Chapter 31 (AI Strategy for the C-Suite), identify the hospital's most promising AI opportunities across three categories: clinical, operational, and administrative. Provide at least two specific use cases in each category.

(b) For your highest-priority clinical use case, describe the data pipeline from raw clinical data to model deployment. Identify at least three points in the pipeline where HIPAA compliance requirements affect your design decisions.

(c) A radiologist pushes back on the proposal to deploy an AI diagnostic assist tool: "I have twenty years of experience. I don't need a machine to tell me what I'm looking at." Using the change management frameworks from Chapter 35, outline a strategy for physician engagement that respects clinical expertise while demonstrating the value of AI augmentation.

(d) NK observes that retail data quality problems are modest compared to healthcare. Describe three specific data quality challenges that are unique to healthcare (not present in retail) and explain how each affects AI model development.

Exercise 8: Manufacturing AI Business Case

A contract manufacturer of automotive components operates five plants with a combined 3,000 employees. The company has no AI capability but has invested in IoT sensors on major equipment over the past two years.

(a) Build a business case for predictive maintenance at one plant. Include: the problem statement, the data requirements, the expected benefits (quantified with reasonable assumptions), the estimated costs, and the implementation timeline.

(b) The plant manager is skeptical: "We've maintained this equipment the same way for thirty years. It works fine." Using the change management approaches from Chapter 35, describe how you would address this resistance without dismissing the plant manager's expertise.

(c) The company is considering two vendors for its predictive maintenance platform: a large industrial AI company (Siemens, GE) and a startup with a more modern technology stack but less manufacturing domain expertise. Apply the build-vs-buy framework from this textbook to analyze this decision. What factors should determine the choice?

(d) After successfully deploying predictive maintenance, what should be the company's second AI use case? Justify your recommendation based on the "crawl-walk-run" progression discussed in this chapter.

Exercise 9: Public Sector Ethical Analysis

A city government is considering deploying an AI system for each of the following applications. For each, identify the primary benefit, the primary risk, the most important ethical consideration, and one governance mechanism that should be in place before deployment.

(a) Predictive policing to allocate patrol resources across the city's 20 police districts.

(b) AI-powered facial recognition at the city's public transit system to identify wanted individuals.

(c) An ML model to predict which families receiving public benefits are committing fraud.

(d) AI-optimized traffic signal timing to reduce congestion and emissions.

(e) An LLM-powered chatbot to handle routine constituent inquiries (building permits, parking tickets, utility billing).

(f) A student success prediction model to identify at-risk students in the city's public school system.

For each application, assess where it falls on the EU AI Act's risk classification (minimal, limited, high, unacceptable) and explain your reasoning.

Exercise 10: Agriculture vs. Energy Comparison

Tom and another student compare their industry assignments — agriculture and energy.

(a) Identify three AI applications that are structurally similar across agriculture and energy. For each pair, describe the parallel and explain what differs (data types, regulatory requirements, deployment context).

(b) Both agriculture and energy face the challenge of AI adoption in rural or remote environments. What infrastructure barriers are common to both industries? What solutions have been proposed or implemented?

(c) Both industries have significant environmental implications. For agriculture, AI can optimize for yield or for sustainability — and these objectives sometimes conflict. For energy, AI can optimize for cost or for carbon reduction — and these objectives sometimes conflict. Choose one industry and describe a specific scenario where the AI optimization objective creates an environmental tradeoff. How should the tradeoff be resolved?

(d) Tom writes: "Agriculture is where energy was ten years ago in terms of AI maturity." Evaluate this claim. In what ways is it accurate? In what ways is it misleading?

Section C: Strategic Analysis (Exercises 11-15)

Exercise 11: The Speed-Governance Tradeoff

The Athena-NovaMart case presents a central strategic tension: NovaMart deploys AI faster by skipping governance, while Athena moves more carefully but risks falling behind competitively.

(a) Map this tradeoff on a 2x2 matrix with "Deployment Speed" on one axis and "Governance Rigor" on the other. Place Athena, NovaMart, Costco (minimal AI), and Amazon (aggressive AI with significant governance) on the matrix. Describe the strategic position and risk profile of each quadrant.

(b) Ravi says, "Speed without responsibility is a liability." Under what conditions is this statement most true? Under what conditions might it be less true? Consider industry, competitive dynamics, and regulatory environment.

(c) Propose a governance process for Athena that maintains rigor while reducing the time from AI project proposal to deployment. Be specific: How long should each governance step take? Who has approval authority? What can be fast-tracked and what cannot?

(d) NovaMart's three lawsuits include two for discriminatory pricing and one for employee surveillance. For each, describe the likely legal theory the plaintiffs are using and assess the strength of the case based on concepts from Chapters 25-30.

Exercise 12: Industry Entry Strategy

You are an AI product company that has built a successful fraud detection platform for financial services. Your board wants you to expand into a second industry.

(a) Evaluate three candidate industries (healthcare, manufacturing, retail) as potential expansion targets. For each, assess the fit of your existing technology, the data requirements, the go-to-market strategy, the competitive landscape, and the regulatory barriers.

(b) Your CTO argues that the core ML technology is identical across industries — only the data and features change. Your VP of Sales argues that industry-specific domain expertise is essential and you need to hire industry veterans. Who is more correct? Build a case for a balanced approach.

(c) Professor Okonkwo says, "The frameworks transfer. The details don't." What specific "details" would you need to learn before deploying your fraud detection technology in healthcare? List at least eight industry-specific factors.

(d) Would you enter the new industry through direct sales to enterprises, through partnerships with industry-specific software vendors, or through acquisition of an industry-specific AI startup? Evaluate each approach with pros and cons.

Exercise 13: Failure Mode Analysis

Select three failure modes from the cross-industry failure modes table in this chapter (pilot purgatory, data debt, governance theater, vendor dependency, talent hoarding, ethics washing).

For each selected failure mode:

(a) Describe a specific, realistic scenario in which this failure mode occurs. Include the industry, the AI application, the organizational context, and the sequence of events that leads to failure.

(b) Identify the early warning signs that would indicate this failure mode is developing. What metrics or observations would an AI leader monitor?

(c) Propose a preventive intervention — something an organization could do before the failure mode manifests to reduce the risk.

(d) Propose a corrective intervention — something an organization could do after the failure mode has been identified to recover.

Exercise 14: Regulatory Impact Comparison

Using Lena's regulatory maturity framework, conduct the following analysis:

(a) Select one industry from each regulatory tier (heavily regulated with AI-specific rules, heavily regulated with nascent AI rules, moderately regulated, lightly regulated). For each, identify a specific AI application and describe how the regulatory environment affects its design, deployment, and ongoing operation.

(b) The EU AI Act classifies AI systems by risk level (unacceptable, high, limited, minimal). Select one AI application from each of three different industries and classify it under the EU AI Act framework. Explain your classification and identify the specific compliance requirements that would apply.

(c) A CEO says, "We operate in a lightly regulated industry, so AI governance is optional." Write a three-paragraph response that explains why governance is important even in the absence of regulation, using specific examples from this chapter.

(d) Predict how AI regulation will evolve over the next five years for one industry of your choice. What regulations are likely to be adopted? How will they affect AI adoption? Will they accelerate or decelerate AI deployment?

Exercise 15: The Capstone Preview

This exercise previews Chapter 39's capstone project. Select an industry not assigned to your team during Professor Okonkwo's in-class exercise.

(a) Identify the three highest-value AI opportunities in that industry, ranked by potential business impact. For each, specify the underlying AI capability, the data requirements, and the expected timeline to value.

(b) Identify the three most significant barriers to AI adoption in that industry, ranked by severity. For each, propose a mitigation strategy.

(c) Identify the three most important ethical considerations for AI in that industry. For each, describe the risk scenario and propose a governance mechanism.

(d) Write a one-page executive summary (approximately 400 words) making the case for AI investment in this industry. Target your audience as the CEO of a mid-size company in the industry who is skeptical of AI but open to evidence.

(e) Professor Okonkwo asks each team to identify one insight from their assigned industry that should be transferred to Athena Retail Group. What is your insight, and how would you implement it at Athena?

Section D: Competitive Intelligence (Exercises 16-18)

Exercise 16: NovaMart Competitive Analysis

NK's competitive analysis of NovaMart reveals a digitally native retailer that deploys AI aggressively across its operations. Based on the information in this chapter, complete the following:

(a) Build a competitive intelligence profile of NovaMart. Infer its likely AI capabilities across five functions: pricing, personalization, supply chain, loss prevention, and customer service. For each function, describe the likely AI approach and the competitive advantage it creates.

(b) Assess NovaMart's strategic vulnerabilities. Identify at least four risks that NovaMart faces as a result of its aggressive, low-governance AI deployment approach.

(c) Draft a briefing for Athena's board (300-400 words) that presents the NovaMart competitive threat while arguing against a "copy NovaMart" strategy. Include specific recommendations for how Athena can compete on AI capability without compromising its governance standards.

(d) Ravi says, "NovaMart moves faster. But they have no governance, no ethics review, and three pending lawsuits. Speed without responsibility is a liability." A board member responds, "But NovaMart's revenue is growing 40% year-over-year while ours grows 6%. Maybe responsibility is the liability." Write Ravi's response (3-4 sentences).

Exercise 17: Industry Competitive Dynamics

Select an industry from this chapter (other than retail).

(a) Identify the industry leader in AI adoption and describe three specific AI capabilities that differentiate them from competitors.

(b) Identify a "fast follower" in the same industry that is rapidly closing the AI capability gap. What strategy are they using?

(c) Identify an incumbent in the industry that has been slow to adopt AI. What are the consequences of their delay? Is there a path to recovery?

(d) Assess whether AI creates winner-take-all dynamics in this industry or whether multiple competitors can sustain AI-driven advantages. Justify your assessment using concepts from competitive strategy.

Exercise 18: Cross-Industry Disruption

AI enables companies to enter industries they previously could not. Google entered healthcare (DeepMind Health). Amazon entered logistics and cloud computing. Tesla entered energy management.

(a) Identify three potential cross-industry disruptions enabled by AI — scenarios where a company from one industry could use AI to enter and disrupt another industry. For each, describe the AI capability that enables the entry, the incumbent vulnerability being exploited, and the likely competitive response.

(b) For one of your scenarios, assess the barriers to entry. Are they primarily technical (data, models, infrastructure), regulatory (licensing, compliance, certification), or organizational (domain expertise, relationships, brand trust)?

(c) From the perspective of an incumbent in the industry being disrupted, propose a defensive strategy that leverages your industry-specific assets (data, customer relationships, regulatory knowledge) against the disruptor's AI capabilities.

Submission Guidelines

Section A: Written responses, 2-4 sentences per sub-question unless otherwise specified. Focus on cross-industry pattern recognition.
Section B: Industry-specific analysis requiring research beyond the chapter content. Use the frameworks from this chapter and earlier chapters as analytical tools.
Section C: Strategic analysis requiring synthesis of concepts from multiple chapters. Longer-form responses expected (one to two paragraphs per sub-question).
Section D: Competitive analysis and strategic reasoning. Support arguments with evidence from the chapter and from your own research.