Skip to content
AAIR Review ManualChapter 1 › Case Study 14 / 33

Case Study

Marmot Home Security is a mid-market enterprise with 3,000 employees and a customer base of over two million worldwide. Marmot develops and sells innovative smart home products including smart speakers, security cameras, and automated lighting systems. The company operates globally with direct-to-consumer sales through e-commerce channels and partnerships with large retailers.

Marmot’s customer service department consists of 150 agents distributed across three regional centers. The service team handles customer inquiries, ranging from technical troubleshooting, warranty service requests, product setup assistance, return processing, and general product information. Customer service agents respond to requests primarily through phone, but they also respond via email, chat, and on social media.

After the release of its complex smart home suite of products, volume grew by 30%, leaving customer service agents overwhelmed. Issues range from high wait times and inconsistent responses to agent burnout.

To address these issues and maintain a competitive edge among industry peers, leadership at Marmot has set a strategic objective to enhance customer experience and operational efficiency via automation and intelligent support systems. However, as many members of the leadership team come from security and insurance backgrounds, they are wary of the privacy and accuracy issues caused by many AI systems.

  1. The head of the customer support team has consulted with the chief technology officer (CTO) on options for leveraging AI and has decided to propose a suite of AI agents to help handle initial customer contact for phone and web chat inquiries. What are the potential benefits they expect to see with the use of AI agents in this case? (Select all that apply.)
    1. Reduce customer wait times by enabling customers to self-serve basic troubleshooting issues.
    2. Triage requests so that high-priority issues are addressed first.
    3. Reduce human resource costs to offset expenses incurred by technology overhead.
    4. Assist human agents in real time with suggestions, knowledge base articles, and sentiment analysis.
  2. The use case has been provided to the risk manager for feedback. Which of the following options should Marmont’s risk manager consider to ensure related risk is addressed? (Select all that apply.)
    1. Ensure AI solutions comply with data privacy regulations specific to regions where Marmot operates.
    2. Evaluate whether AI can accurately understand and process diverse customer inquiries without bias.
    3. Assess the potential impact of AI on reducing customer service agents’ burnout.
    4. Evaluate all AI technologies that are already implemented by competitors.
  3. Which of the following would be the MOST effective approach for gaining senior leadership buy-in at Marmot to develop a risk-aware AI culture, given their concerns related to AI?
    1. Highlight AI’s ability to enhance innovation at any cost.
    2. Outline compliance and model transparency efforts.
    3. Limit AI discussions to align with short-term profitability goals.
    4. Review AI solutions without detailed explanations to avoid technical complexity.
  4. Considering the concerns of Marmot’s leadership about privacy and accuracy issues, what steps can the company take to address ethical issues related to the use of AI agents in their customer service operations, ensuring that solutions maintain customer trust and data integrity?

Chapter 1 Answer Key — Case Study

    1. AI agents can reduce customer wait times with automated responses that link or provide customers with information related to basic troubleshooting issues, enabling low-complexity issues to be resolved by the customer without needing a human service agent.

    2. AI agents can ask customers a series of questions that help them to understand the complexity or nature of the call, enabling them to triage inquiries based on priority or type of support needed. This can ensure that more complex issues are routed to the right teams, reducing the number of human agents that need to engage with a customer before they get to the right person to respond to their issue.

    3. While the use of AI agents may result in a reduction in the workforce or reassignment of personnel to different roles, that is not the main benefit of using AI agents for customer support.

    4. AI agents can be leveraged to provide service agents with real-time knowledge, such as inventory, resolution suggestions, and customer information. This can make responding to inquiries more streamlined for agents, enabling quicker resolution.

    1. Ensuring AI solutions comply with data privacy regulations is crucial for operating globally and maintaining legal and ethical standards.

    2. Evaluating whether AI can accurately understand and process diverse customer inquiries without bias ensures fairness and effectiveness in customer interaction.

    3. Assessing the potential impact of AI on reducing customer service agents’ burnout addresses workforce well-being and operational efficiency.

    4. Focusing solely on AI technologies already implemented by competitors may not address Marmot’s unique customer service challenges and could lead to missed opportunities for innovation.

    1. While highlighting the ways in which AI use enhances and enables innovation can be a key consideration in leveraging the technology, this would not address leadership concerns around security and privacy.

    2. Demonstrating AI’s alignment with compliance and data protection standards is likely to resonate with leaders from security and insurance backgrounds, addressing their concerns about privacy and accuracy while reinforcing the importance of maintaining customer trust. Model transparency will also provide assurance that consideration has been taken to explain how the model makes decisions.

    3. While profitability and ROI are key AI governance considerations, they do not address privacy or security concerns.

    4. Reports and communication related to AI should be tailored to each specific audience, and technical jargon should be avoided to ensure understanding. However, this does not ensure leadership understands how privacy and security issues will be accounted for in the AI development and implementation.

  1. To address ethical issues in using AI agents, Marmot should start by implementing robust data privacy practices that adhere to global data protection laws, ensuring that customer information is securely stored and processed. Marmot can use AI algorithms that are transparent and explainable, allowing both customers and service agents to understand how decisions are made. Regular audits of AI performance for bias and accuracy should be conducted to maintain fairness. Engaging customers in feedback loops will also help refine AI interactions and build trust. Additionally, continuing education and training for employees about the ethical use of AI can foster an enterprisewide understanding and commitment to RAI deployment.

Footnotes

  1. Andreessen Horowitz, “AI Glossary by a16z,” https://a16z.com/ai-glossary/
  2. Thakur, D.; Saini, J.K.; et al.; “DeepThink IoT: The Strength of Deep Learning in Internet of Things,” Artificial Intelligence Review, vol. 56, 2023, p. 14663-14730, https://doi.org/10.1007/s10462-023-10513-4
  3. Dutta, B.; “7 Types of Artificial Intelligence You Should Know,” Analytics Steps, 31 May 2022, link
  4. White, C.; “What are Reactive Machines in Artificial Intelligence? Meaning, Examples and Uses,” Articlesbase.com, 27 April 2024, link
  5. Muns, A.; “Key characteristics of limited memory AI systems,” Telnyx, https://telnyx.com/learn-ai/limited-memory-ai
  6. Cuzzolin, F.; Morelli, A.; et al.; “Knowing me, knowing you: theory of mind in AI,” Psychological Medicine, vol. 50, iss. 7, 2020, p. 1057-1061, link
  7. Most Inside, “What Are the Different Types of AI? What Are the Career Options in AI?,” 14 October 2020, link
  8. Investopedia, “Weak AI (Artificial Intelligence): Examples and Limitations,” 1 January 2022, link
  9. Kanade, V.; “What Is Narrow Artificial Intelligence (AI)? Definition, Challenges, and Best Practices for 2022,” Spiceworks, 22 March 2022, link
  10. Deb, S.; “Artificial Intelligence Tutorial—A Comprehensive Guide To AI,” Edureka, 1 March 2019, link
  11. GeeksforGeeks, “What Is Artificial Narrow Intelligence (ANI)?,” 15 July 2025, link
  12. Fitzsimmons, M.; Hicks, M.; “Self-driving cars: your complete guide to autonomous vehicles,” TechRadar, 7 June 2019, link
  13. Gupta, M.; “Artificial General Intelligence (AGI) – Definition, Examples, Challenges,” GeeksforGeeks, 15 July 2025, link
  14. Berruti, F.; Nel, P.; et al.; “An executive primer on artificial general intelligence,” McKinsey & Company, 29 April 2020, link
  15. Berruti, et al., “An executive primer”
  16. Nash, C.; “Create Data from Random Noise With Generative Adversarial Networks,” Toptal, link
  17. Purdy, M.; “What Is Agentic AI, and How Will It Change Work?,” Harvard Business Review, 12 December 2024, link; Finn, T.; Downie, A.; “Agentic AI vs. generative AI,” IBM, 11 February 2025, link
  18. Purdy, “What Is Agentic AI”; Finn, et al., “Agentic AI vs. generative AI”
  19. Purdy, “What Is Agentic AI”
  20. Murali, A.; “Safeguarding the Enterprise AI Evolution: Best Practices for Agentic AI Workflows,” ISACA Industry News, 1 July 2025, link
  21. Blackman, R.; “Organizations Aren’t Ready for the Risks of Agentic AI,” Harvard Business Review, 13 June 2025, link
  22. Blackman, “Organization’s Aren’t Ready”
  23. Kurama, V.; “Regression in Machine Learning: Definition and Examples of Different Models,” Built In, 1 August 2024, link
  24. Shetty, B.; “5 Classification Algorithms for Machine Learning,” Built In, 12 April 2023, link
  25. IBM, “What is unsupervised learning?,” link; Baheti, P.; “Supervised vs. Unsupervised Learning [Differences & Examples],” V7 Labs, 1 October 2021, link
  26. Google for Developers, “Clustering Algorithms,” 18 July 2022, link; Avarind, C.R.; “Exploring Clustering Algorithms: Explanation and Use Cases,” Neptune.ai, 9 August 2023, link; Alboukadel; “Types of Clustering Methods: Overview and Quick Start R Code,” Datanovia, 4 November 2018, link
  27. De Vries, T.; “Anomaly Detection in IT Audit,” University of Turku, 31 May 2022, link; IBM, “What is unsupervised learning?”
  28. Dobilas, S.; “Reinforcement Learning (RL) — What Is It and How Does It Work?,” Towards Data Science, 26 September 2022, link
  29. Dickson, B.; “What is neural architecture search?,” TechTalks, 28 February 2022, link; Zoph, B.; Le, Q.; “Neural Architecture Search with Reinforcement Learning,” link
  30. Elsken, T.; Hendrik Metzen, J.; et al.; “Neural Architecture Search: A Survey,” Journal of Machine Learning Research, vol. 20, 2019, p. 1-21
  31. Dickson, “What is neural architecture search?”
  32. Cambridge Dictionary, “Algorithm,” link
  33. SplashLearn, “Algorithm in Math – Definition with Examples,” link
  34. IBM, “What is unsupervised learning?”
  35. Sarah, M.; “A Comprehensive Guide to Cluster Analysis: Applications, Best Practices and Resources,” Displayr, link
  36. Cheng, C.; “Principal Component Analysis (PCA) Explained Visually with Zero Math,” Towards Data Science, 3 February 2022, link
  37. Awati, R.; Burns, E.; “Ensemble modeling,” TechTarget, 9 May 2024, link; Naderalvojoud, B.; Hernandez-Boussard, T.; “Improving machine learning with ensemble learning on observational healthcare data,” AMIA Annual Symposium Proceedings Archive, 2023, p. 521-529, link
  38. Awati and Burns, “Ensemble modeling”; Murel, J.; Kavlakoglu, E.; “What is ensemble learning?,” IBM, 18 March 2024, link
  39. AWS, “What is LLM (Large Language Model)?,” link; Turing, “Understanding Transformer Neural Network Model in Deep Learning and NLP,” link
  40. Suiche, M.; “Building Agents for Small Language Models: A Deep Dive into Lightweight AI,” 27 August 2025, link; GeeksForGeeks, “What are Small Language Models (SLMs),” 23 July 2025, link
  41. Georgetown University, “Artificial Intelligence (Generative) Resources,” link
  42. AWS, “What are Foundation Models?,” link; IBM, “What are foundation models?,” 9 May 2022, link
  43. Chen, S.; Lin, Z.; et al.; “Critical Attention Scaling in Long-Context Transformers,” 7 October 2025, https://arxiv.org/pdf/2510.05554
  44. AWS, “What is Natural Language Processing (NLP)?,” link; IBM, “What is NLP (natural language processing)?,” link
  1. IBM, “What is sentiment analysis?,” link
  2. IBM, “What is underfitting?,” link
  3. GenAI for California, “Evaluating use cases,” link; ISACA, ISACA AAISM Official Review Manual, USA, 2025
  4. European Commission, “European approach to artificial intelligence,” link
  5. Larson, B.; Dignum, V.; “AI value alignment: How can we align artificial intelligence with human values,” World Economic Forum, 17 October 2024, link
  6. World Economic Forum, “AI Value Alignment: Guiding Artificial Intelligence Towards Shared Human Goals,” October 2024, link
  7. Morthorst, T.; “How organizations can choose between buying and building AI systems,” EY, 19 February 2024, link
  8. US National Institute of Standards and Technology (NIST), Artificial Intelligence Risk Management Framework (AI RMF), USA, 2023, link
  9. ISACA, Leveraging COBIT for Effective AI System Governance, 31 January 2025, link
  10. Committee of Sponsoring Organizations of the Treadway Commission (COSO), Enterprise Risk Management—Integrating with Strategy and Performance Framework, USA, 2017
  11. COSO, Applying the COSO Framework and Principles to Help Implement and Scale Artificial Intelligence, USA, 2021
  12. Ryseff, J.; De Bruhl, B.F.; et al.; “The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed,” Rand, 13 August 2024, link
  13. De Marco, N.; “Building Resilient Organizations With AI,” Forbes, 29 September 2023, link
  14. Scarpino, J.; “Ethics, Accountability, and the Pursuit of Responsible AI,” ISACA Journal, vol. 3, 2025, link
  15. Scarpino, “Ethics, Accountability”
  16. ISACA, Leveraging COBIT
  17. ISACA, The Promise and Peril of the AI Revolution: Managing Risk, USA, 2023
  18. US NIST, AI RMF
  19. US NIST, AI RMF Playbook, link
  20. Howard, M.; Martucci-Moore, S.; “Artificial Intelligence Doesn’t Own Its Outputs, so Who Does? AI Ownership and Related Contractual Considerations,” Krevolin & Horst, 13 November 2024, link
  21. Schmidt, S.; “Bartz v. Anthropic: Settlement reached after landmark summary judgment and class certification,” Inside Tech Law, 5 September 2025, link
  22. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), link
  23. United Nations Educational, Scientific, and Cultural Organization (UNESCO), “Ethical Impact Assessment,” link
  24. Scarpino, J.; “Deploying Responsible AI,” ISACA Journal, vol. 1, 2024, link
  25. Schwartz, R.; Vassilev, A.; et al.; NIST Special Publication 1270, Towards a Standard for Identifying and Managing Bias in Artificial Intelligence, USA, March 2022, link
  26. Nkonde, M.; “Is AI Bias a Corporate Social Responsibility Issue?,” Harvard Business Review, 4 November 2019, link
  27. Hardesty, L.; “Study finds gender and skin-type bias in commercial artificial-intelligence systems,” MIT News, 11 February 2018, link
  28. The Linux Foundation, “AI Fairness 360: Understand and mitigate bias in ML models,” https://ai-fairness-360.org/
  29. Wexler, J.; Pushkarna, M.; et al.; “The What-If Tool: Interactive Probing of Machine Learning Models,” link
  30. Jonker, A.; Gomstyn, A.; et al.; “What is AI transparency?,” IBM, 6 September 2024, link
  31. Putrus, R.; “AI-Infused Foresight: Unveiling Megatrends for Strategic Adaptation,” ISACA Journal, vol 3., 2024, link
  32. Merica, D.; Bedayn, J.; “As data centers proliferate, conflict with local communities follow,” The Associated Press, 5 December 2024, link
  33. Merica and Bedayn, “As data centers proliferate”
  34. Brynjolfsson, E.; Chandar, B.; et al.; “Canaries in a Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence,” Stanford Digital Economy Lab, August 2025, link
  35. International Energy Agency (IEA), “Electricity 2024,” link
  36. Marrinan, C.; “Data Center Boom Risks Health of Already Vulnerable Communities,” Tech Policy, 12 June 2025, link
  37. James, L.; “AI data centers are swallowing the world’s memory and storage supply, setting the stage for pricing apocalypse that could last a decade,” MSN, link
  38. IEA, “Electricity 2024”
  39. UNESCO, “Recommendation on the ethics of artificial intelligence,” 2021, link
  40. Marrinan, C.; “Data Center Boom Risks”
  41. NVIDIA, “NVIDIA Introduces Spectrum-XGS Ethernet to Connect Distributed Data Centers into Giga-Scale AI Super-Factories,” 22 August 2025, link
  42. European Space Agency (ESA), “Artificial intelligence in space,” link
  43. Ennamli, A.; “From Risk to Reward: Harnessing AI for Environmental Sustainability,” ISACA Journal, vol. 4, 2025, link