Chapter 34 Further Reading: Measuring AI ROI

AI ROI Measurement and Frameworks

  1. Bughin, J., Seong, J., Manyika, J., Chui, M., & Joshi, R. (2018). "Notes from the AI Frontier: Modeling the Impact of AI on the World Economy." McKinsey Global Institute Discussion Paper. A large-scale analysis of AI's economic impact across industries and geographies. Provides the macro-level context for AI ROI discussions, including estimates of AI's potential contribution to global GDP ($13 trillion by 2030). Useful for framing AI investments in terms of industry-wide opportunity and for executive presentations that require broad economic context.

  2. McKinsey & Company. (2024). "The State of AI in Early 2024: Gen AI Adoption Spikes and Starts to Generate Value." McKinsey Global Survey. The most recent installment of McKinsey's annual AI survey, covering AI adoption, ROI, organizational practices, and the emerging impact of generative AI. The data on AI high performers — including the finding that 25 percent of respondents now report >20% of EBIT from AI — is the foundation of Case Study 1. Essential for benchmarking your organization's AI maturity and ROI.

  3. Ransbotham, S., Khodabandeh, S., Kiron, D., Candelon, F., Chu, M., & LaFountain, B. (2020). "Expanding AI's Impact With Organizational Learning." MIT Sloan Management Review and Boston Consulting Group Report. Based on a survey of over 3,000 managers from 112 countries, this report examines why most organizations fail to generate value from AI. Finds that organizational learning — the ability to absorb lessons from AI projects and apply them systematically — is the strongest predictor of AI success. Complements the McKinsey research with a deeper focus on the learning mechanisms that drive ROI.

  4. Davenport, T.H. & Ronanki, R. (2018). "Artificial Intelligence for the Real World." Harvard Business Review, 96(1), 108-116. Based on a study of 152 AI projects, categorizes AI applications into three types: process automation, cognitive insight, and cognitive engagement. Each type has a different ROI profile. Process automation delivers quick, measurable returns; cognitive insight creates analytical value; cognitive engagement transforms customer interactions. Originally cited in Chapter 6's further reading — revisit with the ROI lens from this chapter.

  5. Bessen, J. & Righi, C. (2019). "Shocking Technology: What Happens When Firms Make Large IT Investments?" Boston University School of Law, Law and Economics Paper No. 19-6. Examines the financial trajectory of firms making large IT and AI investments, using firm-level data. Finds the J-curve pattern described in Section 34.6: large technology investments typically depress performance for 3-5 years before generating measurable returns. Critical reading for anyone managing expectations during the investment phase of an AI program.

Financial Analysis for Technology Investments

  1. Brealey, R.A., Myers, S.C., & Allen, F. (2020). Principles of Corporate Finance, 13th Edition. McGraw-Hill. The standard MBA textbook on corporate finance, covering NPV, IRR, cost of capital, and real options. The NPV and IRR calculations in the AIROICalculator are drawn directly from the frameworks in this text. For readers who want a deeper understanding of the financial theory underlying AI ROI analysis, Chapters 5-6 (on NPV and project evaluation) and Chapter 22 (on real options) are particularly relevant.

  2. Dixit, A.K. & Pindyck, R.S. (1994). Investment Under Uncertainty. Princeton University Press. The foundational academic treatment of real options theory — the framework used in Section 34.5 to analyze strategic optionality. Demonstrates how traditional NPV analysis undervalues investments with significant uncertainty and flexibility. Dense and mathematical, but the introductory chapters are accessible to MBA readers and provide the intellectual foundation for option-value thinking in AI investment.

  3. Luehrman, T.A. (1998). "Investment Opportunities as Real Options: Getting Started on the Numbers." Harvard Business Review, July-August 1998. A practitioner-friendly introduction to real options analysis for business investments. Provides a bridge between the abstract theory of Dixit and Pindyck and the practical application to technology investments. Directly relevant to the option value estimation framework in Section 34.5.

AI Project Portfolio Management

  1. Fountaine, T., McCarthy, B., & Saleh, T. (2019). "Building the AI-Powered Organization." Harvard Business Review, 97(4), 62-73. Three McKinsey partners describe the organizational requirements for AI at scale, based on work with hundreds of companies. Key insight: 80 percent of the work required to capture value from AI is organizational, not technical. Covers AI portfolio management, talent strategy, and the organizational changes required to move from pilot to production. Directly relevant to the portfolio management framework in Section 34.8.

  2. Ng, A. (2018). "AI Transformation Playbook." Landing AI White Paper. Andrew Ng's concise guide to enterprise AI transformation, including a framework for selecting and sequencing AI projects. Recommends starting with quick wins (0-6 months), building a team and infrastructure (6-18 months), and then pursuing strategic projects (12-36 months). The sequencing logic aligns with the portfolio balance recommendations in Section 34.8. Also cited in Chapter 6; revisit with the ROI measurement lens.

  3. Iansiti, M. & Lakhani, K.R. (2020). Competing in the Age of AI. Harvard Business Review Press. Examines how AI-native companies build competitive advantages through data and algorithms, creating "operating models" that scale in ways traditional businesses cannot. The competitive positioning analysis in Section 34.5 draws on the concepts developed here. Particularly relevant for understanding why some AI investments create compounding returns while others generate diminishing ones.

Total Cost of Ownership and ML Economics

  1. Sculley, D., et al. (2015). "Hidden Technical Debt in Machine Learning Systems." Advances in Neural Information Processing Systems (NeurIPS), 28. The seminal paper on the ongoing costs of ML systems in production, originally cited in Chapter 6. Argues that ML systems accumulate technical debt faster than traditional software due to entanglement, hidden feedback loops, undeclared consumers, and data dependencies. Provides the intellectual foundation for the TCO analysis in Section 34.9 — particularly the finding that operations costs dominate the lifecycle.

  2. Paleyes, A., Urma, R.G., & Lawrence, N.D. (2022). "Challenges in Deploying Machine Learning: A Survey of Case Studies." ACM Computing Surveys, 55(6), 1-29. A survey of 99 published ML deployment case studies, documenting the real-world costs and challenges of putting models into production. Finds that organizational and infrastructure challenges are cited more frequently than modeling challenges. Empirically validates the hidden cost analysis in Section 34.3.

  3. Polyzotis, N., Roy, S., Whang, S.E., & Zinkevich, M. (2018). "Data Lifecycle Challenges in Production Machine Learning." ACM SIGMOD Record, 47(2), 17-28. Google researchers describe the data management costs in production ML pipelines. Data costs — acquisition, preparation, validation, and maintenance — often exceed compute and modeling costs combined, yet are the most frequently underestimated cost category. Directly relevant to the data costs section of the AI Cost Taxonomy (Section 34.3).

Autonomous Vehicles and Moonshot ROI

  1. Higgins, T. (2021). Power Play: Tesla, Elon Musk, and the Bet of the Century. Doubleday. While focused on Tesla rather than Waymo, this book provides essential context for Case Study 2 by examining the competitive dynamics of autonomous vehicle development. The contrast between Tesla's camera-only approach and Waymo's LIDAR-based approach illustrates the strategic uncertainty inherent in moonshot AI investments.

  2. Burns, L.D. & Shulgan, C. (2018). Autonomy: The Quest to Build the Driverless Car — And How It Will Reshape Our World. Ecco/HarperCollins. A history of autonomous vehicle development by a former GM VP of R&D. Covers the DARPA Challenges, the Google Chauffeur project, and the early competitive landscape. Provides the historical context for Waymo's investment trajectory and the extraordinary time horizons involved in moonshot AI.

  3. Krafcik, J.F. (2020). "Waymo's Approach to Safety." Waymo Safety Report (3rd Edition). Waymo's own safety framework document, detailing its approach to validating autonomous driving safety. Illustrates the governance and compliance costs associated with safety-critical AI systems — costs that are often excluded from traditional ROI analyses but can be substantial. Available from Waymo's website.

Behavioral Economics and Decision-Making

  1. Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux. The definitive popular treatment of cognitive biases, including the sunk cost fallacy, optimism bias, and planning fallacy — all of which influence AI investment decisions. The discussion of kill criteria in Section 34.7 draws directly on Kahneman's work on how pre-committed decision rules can counteract cognitive biases. Essential reading for anyone who makes or evaluates AI investment decisions.

  2. Arkes, H.R. & Blumer, C. (1985). "The Psychology of Sunk Cost." Organizational Behavior and Human Decision Processes, 35(1), 124-140. The foundational empirical study of the sunk cost fallacy, demonstrating through controlled experiments that people (including trained economists and business professionals) systematically factor sunk costs into forward-looking decisions. Directly relevant to the kill-or-continue decisions discussed in Section 34.7.

Communication and Executive Reporting

  1. Duarte, N. (2010). Resonate: Present Visual Stories That Transform Audiences. Wiley. A guide to building presentations that engage and persuade. The narrative strategies in Section 34.11 — customer stories, counterfactuals, and competitive framing — are applications of the storytelling principles Duarte describes. Useful for anyone preparing an AI ROI presentation for executive or board audiences.

  2. Zelazny, G. (2001). Say It With Charts: The Executive's Guide to Visual Communication, 4th Edition. McGraw-Hill. The classic reference on data visualization for business communication. Covers the selection and design of charts for different types of data and messages. Relevant to the dashboard design principles in Section 34.11, particularly the design of the three-layer reporting structure (portfolio summary, project scorecards, detailed analysis).

Industry Reports and Benchmarks

  1. Gartner. (2024). "Predicts 2025: AI's Impact on Business Will Accelerate Despite Challenges." Gartner's annual predictions report on AI in business, including time-to-value benchmarks, deployment success rates, and organizational maturity models. The time-to-value table in Section 34.6 draws on Gartner's benchmarking data. Available through Gartner's subscription service or summarized in public press releases.

  2. IDC. (2023). "Worldwide Artificial Intelligence Spending Guide." IDC's comprehensive analysis of global AI spending by industry, technology, and use case. Includes ROI estimates across industries and company sizes. The "$3.50 per $1 invested" median figure cited in Section 34.12 is drawn from IDC's analysis. Useful for benchmarking your organization's AI spending and returns against industry averages.

  3. Accenture. (2024). "The Art of AI Maturity: Advancing from Practice to Performance." Based on a survey of over 1,600 organizations, this report identifies the practices that separate AI leaders from laggards — with a particular focus on revenue growth differentials. The 2.5x revenue growth gap cited in Section 34.12 comes from this research. Includes a useful AI maturity assessment framework.

  4. MIT Sloan Management Review. (2023). "Achieving Individual — and Organizational — Value With AI." MIT SMR and BCG Annual AI Report. The annual report from MIT Sloan's AI research initiative, based on surveys and case studies. The finding that only 10 percent of companies report "significant financial benefits" from AI is drawn from this report. Provides a more cautious perspective on AI ROI than the consulting firm reports, grounded in academic research methodology.

These readings span financial theory (Brealey, Dixit and Pindyck), organizational strategy (Fountaine, Iansiti), behavioral economics (Kahneman, Arkes), technical economics (Sculley, Paleyes), and industry research (McKinsey, Gartner, IDC). For readers short on time: start with McKinsey (2024) for benchmarks, Sculley (2015) for hidden costs, and Kahneman (2011) for the cognitive biases that distort every ROI analysis. For readers preparing AI ROI presentations, add Duarte (2010) and Zelazny (2001).